8.30am – 10am
10.30am – 12pm
4pm – 5.30pm
Violet Theater
Recent advances in binding and retrieval in action control I: Action plans, responses and outcomes
SymposiumRecent advances in binding and retrieval in action control I: Action plans, responses and outcomes
Symposium organized by Malte Möller and Mrudula Arunkumar
Humans interact with their environment with remarkable ease. This is astonishing given that, in most situations, humans have the option to act in various ways, but must choose the appropriate actions to achieve a certain goal. Mechanisms underlying human action control have been traditionally studied in several research domains, focussing on different but related phenomena, theories, and experimental approaches. Recently, the Binding and Retrieval in Action Control framework (BRAC, Frings et al. 2020) offered a unifying perspective on research on human action control by proposing stimulus binding and retrieval as two basic, functionally separable processes that can account for a broad range of related experimental phenomena across a variety of tasks. The first part of the symposium introduces the BRAC framework and its core mechanisms, followed by recent advances regarding the continuous measurement of the impact of S-R binding on performance, the formation of S-R episodes in action planning and after error commission as well as the binding between sequential responses. The last contribution of part I is concerned with the impact of affective consequences on binding and retrieval of S-R episodes. Together, the research presented in the symposium highlights the utility of the BRAC framework as a starting point to generate new insights into the mechanisms that guide controlled human behaviour.
8.30am
Introducing BRAC: Current research on binding and retrieval
Andrea M. Philipp, Juhi Parmar, Mrudula Arunkumar, & Malte Möller
One of the core ideas of the BRAC (Binding and Retrieval in Action Control) framework is that action control ‐ in a variety of situations - can be explained by two basic mechanisms: Binding and Retrieval. Binding hereby is the mechanism that combines different elements of a stimulus-response episode into an integrated event-file. When one or more of these elements are repeated later on, this leads to a retrieval of the event-file and, thus, influences performance. In this introductory presentation, we will summarize current research on binding and retrieval and present further developments of the BRAC framework. Different research questions will be addressed, such as which factors can influence binding and retrieval - and which cannot. Further, we also discuss temporal aspects of binding and retrieval as well as the interplay of both mechanisms but also their distinct influences on action control. Finally, we would like to speculate on how broadly the BRAC framework can be defined and whether it may be able to account for empirical phenomena in other research areas like learning, memory, and attention.
When responding to a stimulus, stimulus and response features are integrated into a so-called event file. When some or all features comprised in an event file repeat in a later episode the whole event file is retrieved. When a later episode is identical to the episode in an event-file, performance benefits emerge because the previous response can be recycled without additional computations. When only some features match between previous and present episode, performance costs manifest because the conflict between previous and present response has to be resolved. The combination of these costs and benefits is termed stimulus-response binding effect (S-R binding). Previous research on S-R binding focused on key-press tasks where cost and benefits are measured as a single reaction time for each trial. However, this single value obscures several underling processes that get lost when observing responses as keypresses. Here, we present a novel adaptation of the established distractor-response binding paradigm that uses mouse movements instead of keypresses. This allows to further distinguish between a preparatory and an executive (motor) phase. We present several experiments that used this novel measure technic to delve deeper into how S-R binding is modulated.
9.00am
Binding and retrieval of spatial and temporal features of planned actions
Viola Mocke, Patricia Holzmann, Bernard Hommel, & Wilfried Kunde
According to the BRAC framework, action planning relies on binding codes of the relevant action features. Such binding is indicated by the observation that planning a novel action is impaired if it shares some, but not all, features with another action that is held in memory for later execution. In the typical paradigm, participants plan a rather complex series of keypresses with one hand, but, before executing it, plan and execute a simple keypress with the same (partial feature overlap) or the other hand (no feature overlap). Here, only the features left and right can be shared between the two action plans. Using an online paradigm with simple keypresses only, we were able to replicate these findings of binding effects for horizontal features as well as extend them to temporal (short and long) and vertical (top and bottom) features. Two important observations were made. First, we found a robust effect of uncertainty on binding. Generally, when action features remain uncertain until the stimulus demanding that action appears, then these features have better chances of being bound than when they are known already from the start of the experiment. However, in the latter case, features may still be bound despite their certainty, and whether they are seems to depend on the feature type. Second, we found a robust repulsion effect between the two action plans. Specifically, within the possibilities of the task, the execution of action features was shaped in a way that these features are distinguishable from identical features held in memory for another subsequent action. For instance, with a short keypress planned for later execution, an intermediate keypress will be longer than with a long keypress held in memory.
9.15am
What is left after an error?
Anna Foerster, Birte Moeller, Christian Frings, & Roland Pfister
Agents create short-cuts in action control by binding features of an acted upon stimulus to features of the response. Re-encountering the stimulus retrieves the bound response, facilitating its execution. If agents commit an error, binding steps in as a corrective measure because the intended, not executed correct response instead of the executed erroneous response enters a binding with the presented stimulus. The current study investigated whether this adaptive goal-based binding and retrieval mechanism depends on error awareness, and it scrutinized the fate of the executed erroneous response in the process of binding and retrieving the intended correct response. The results demonstrate similar binding and retrieval of the intended correct response with and without explicit error feedback, pointing to a critical role of internal error signals instead. At the same time, the data still reveal a continued impact of the unbound erroneous response on action control.
According to current action control theories, individual responses are represented via short-term memory units called event files (Hommel et al., 2001; Frings et al., 2020) that have been described as loose networks of bindings between stimulus, response, and effect features (Hommel, 2004). Thus, responses are at least partly represented by (response relevant) stimuli and following effects. Recently, it has been shown that individual responses can also enter bindings with other responses (response-response bindings, Moeller & Frings, 2019). That is, two sequential responses can be integrated so that repeating one of them later on retrieves the other, influencing further responding. These binding effects are smaller for responses executed by different hands that are positioned far apart than by hands positioned close together. Here we analyzed whether the distance between target stimuli or effects of responses affected bindings between the responses in a similar way. In two experiments, using a response-response binding paradigm, we either grouped the effects elicited by to be integrated responses (\(N = 30\)) or the target stimuli that participants responded to (\(N = 40\)). Results indicate that grouping of stimuli and of effects influenced response-response binding differently. Possible indications for the ideomotor principle in the context of binding and retrieval are discussed.
9.45am
Stimulus-response binding and retrieval is independent of affective consequences: Implications for theories of learning and behaviour automatization
Juhi Parmar & Klaus Rothermund
Stimulus-response binding and retrieval (SRBR) is a fundamental mechanism driving behaviour automatization. In 4 experiments, we investigated the modulatory role of performance-dependent and performance-independent affective consequences on stimulus-response binding and retrieval (SRBR) in order to test whether binding/retrieval can explain instrumental learning. SRBR effects were assessed with a colour categorisation task in a sequential prime-probe design, with an orthogonal variation of Response relation (colour repetition vs. change) by Distractor relation (word repetition vs. change). Binding/retrieval effects are measured by an interaction of the two factors, with distractor repetition leading to facilitation in conditions where the response must be repeated, but leading to interference when the response changes from prime to probe. Positive, neutral, or negative events signaling changes in points were delivered after every trial, which were linked to monetary gains or losses on a block level, to investigate whether affective consequences modulate the binding/retrieval effect. Consistently across all 4 experiments (total \(n = 338\)), we did not find any evidence for an affective modulation of RT-based binding/retrieval effects, indicating that these effects are fully automatic and independent of (performance dependent and independent) feedback. Results are discussed with regard to their implications for theories of learning and habit formation.
Recent advances in binding and retrieval in action control II: Context, learning, and binding by observation
SymposiumRecent advances in binding and retrieval in action control II: Context, learning, and binding by observation
Symposium organized by Andrea M. Philipp and Juhi Parmar
Humans interact with their environment with remarkable ease. This is astonishing given that, in most situations, humans have the option to act in various ways, but must choose the appropriate actions to achieve a certain goal. Mechanisms underlying human action control have been traditionally studied in several research domains, focussing on different but related phenomena, theories, and experimental approaches. Recently, the Binding and Retrieval in Action Control framework (BRAC, Frings et al. 2020) offered a unifying perspective on research on human action control by proposing stimulus binding and retrieval as two basic, functionally separable processes that can account for a broad range of related experimental phenomena across a variety of tasks. The second part of the symposium starts with a focus on the inclusion of task-irrelevant cue and context information in S-R bindings, the structure of S-R bindings involving context, as well as the relation between context binding and cognitive control. The subsequent talk connects the BRAC framework to contingency learning while the final contribution presents evidence that S-R bindings can be acquired by observation even in online settings beyond the laboratory. Together, the research presented in the symposium highlights the utility of the BRAC framework as a starting point to generate new insights into the mechanisms that guide controlled human behaviour.
10.30am
Inter-trial variability of context influences the binding structure in a stimulus-response episode
Ruyi Qiu, Malte Möller, Iring Koch, & Susanne Mayr
Several lines of research suggest that stimuli and responses are bound together in a binary fashion into an episodic representation, called event file or stimulus-response episode. However, Mayr et al. (2018) employed a four-alternative negative priming task which required participants to perform a keypress response to a target sound and to ignore a simultaneously presented distractor sound. Interestingly, results showed that a context, that is, a completely task-irrelevant sound, modulated the binary binding between the distractor stimulus and the response, instead of entering into a binary binding with the response itself. The current study investigates the role of inter-trial variability of the context on the binding structure between the context and the response. The study employed the auditory negative priming task previously used by Mayr et al. Participants were either assigned to the high-variability group (with 8 different individual sounds as context) or the low-variability group (with only 2 different individual sounds as context). In the low-variability group, context modulated the binary binding between the distractor stimulus and the response, thereby replicating the previous finding. In the high-variability group, however, the repetition of the context sound per se retrieved the prime response, indicating a binary binding between the context and the response. Together, the current findings provide evidence that the variability property is a determinant of the binding structure involving context information in a stimulus-response episode. Possible underlying mechanisms (i.e., attention and cue overload) are discussed.
10.45am
Binding of task-irrelevant cue features in cued language-switching
Elena Benini, Iring Koch, Susanne Mayr, Christian Frings, & Andrea M. Philipp
The processes underlying language control are often investigated through language switching paradigms. Previous research suggested considerable overlap in control mechanisms in language switching and task switching. Thus, as feature binding and retrieval processes are likely to play a crucial role in task switching performance, we expect that feature binding and retrieval also influences language control. In the present study, we use a picture-naming language-switching paradigm, in which the currently relevant language is indicated by a symbolic cue that is presented either visually or auditorily. In two experiments (\(N = 43\) and \(N = 46\)), we find evidence that the irrelevant cue modality (auditory vs. visual) is bound with the language and retrieves it when the cue modality repeats in the subsequent trial. That is, language repetition benefits are larger when the cue modality repeats than when it switches. Further, in the second experiment, the data indicate a trend towards larger binding and retrieval effects with low-frequent words as compared to high-frequency words. Taken together, these results suggest that binding and retrieval mechanisms interfere with language control processes and that such interference is modulated by the frequency of use (i.e., practice) for each word. We discuss implications of the interaction between language control, language proficiency, and short-term features bindings.
11.00am
Contextualized control ‐ disruption or retrieval?
Moritz Schiltenwolf, Hannah Dames, Jessica Helm, Andrea Kiesel, Christina U. Pfeuffer, & David Dignath
Recent literature shows that cognitive control is dependent on contextual information. For example, it has been shown that the congruency sequence effect (CSE), the most widely used behavioral marker for control adaptations, is larger in trial sequences with identical context in prime and probe than in sequences with context changes between prime and probe. Binding accounts suggest that this is due to control states that become bound to the context, and re-encountering that context leads to the retrieval of the bound control state. However, this view is challenged by a disruption account stating that control is proactively maintained until it is disrupted by change in context. We contrasted, these two accounts by adding a disruption trial between prime and probe displayed in a context that is always different to prime and probe (e. g. A-B-A or A-B-C context sequences). The binding account would predict control state retrieval, if prime and probe trial share the same context. However, the disrupt account would predict a continuous control decrease, as the context changes in all successive trials (prime to disruption trial and disruption trial to probe). I will discuss findings from this paradigm with regard to binding and control theories.
11.15am
Can associated stimuli indirectly retrieve a response from another paired stimulus? A test for transfer of associations and indirect retrieval using a contingency learning task
Mrudula Arunkumar, Klaus Rothermund, & Carina G. Giesen
If a stimulus is contingently paired with a particular response, this impacts on performance and typically yields benefits (costs) for contingent (noncontingent) stimulus-response pairings. Contingency learning can be explained by stimulus-based episodic retrieval of past responses. However, it is yet unresolved whether response retrieval can also be triggered by another stimulus that was never directly paired with the response, but is only associated with the former stimulus. From the classical conditioning literature, this is demonstrated by the concept of sensory preconditioning. We investigated the phenomenon of indirect retrieval in a contingency learning paradigm. Phase I: Two words were associated with each other (S1-S2 pairings). Phase II: S2 was contingently paired with a response (S2-R pairings). Phase III then tested contingency learning for S2 (manipulation check) and S1 (test for indirect retrieval) in a forced-choice task (study 1) and a free-choice task (study 2). We will discuss the findings of our studies, which will further extend our understanding of the cognitive mechanisms underlying learning effects.
11.30am
BRAC goes social: Retrieval of observational stimulus-response bindings in online settings
Carina G. Giesen
Observing how another person responds to a stimulus results in observationally acquired stimulus-response (SR) bindings. These can be retrieved from memory on later occasions, which means that observed responses are utilized for regulating one's own actions. Until now, evidence for storage and retrieval of observationally acquired SR bindings was limited to dyadic interactions between two interaction partners who respond in alternating fashion. In two highly powered, pre-registered studies, I demonstrate for the first time that observational SR bindings can also be acquired in online interactions: Robust retrieval effects emerged when observers believe to be interacting with another person. In turn, retrieval effects were absent when observers believe to be interacting with the computer. The findings show that feature-based binding and retrieval principles are pervasive and also apply to social interactions, regardless of real-life or online.
Action binding
Action binding
Session chaired by Birte Moeller
4.00pm
It takes some time ‐ decay of response-response binding effects during the first twelve seconds after integration
Birte Moeller & Christian Frings
Several influential accounts of human action control assume transient bindings between stimulus- and response features of individual events (SR-binding; e.g., Frings et al., 2020; Hommel et al., 2001; Schmidt et al., 2016). Notably, the duration of bindings seems to increase with the task relevance of the bound features. Recently, we observed relatively long durations of bindings between individually planned and executed responses (RR-binding): In contrast to SR-bindings that typically decay completely within five seconds after integration, we found no measurable decay of RR-bindings within the first six seconds after integration (Moeller & Frings, 2021). Here we aimed to pinpoint a timeframe in that RR-bindings decay. In the presented Experiment (\(N = 29\)) we measured response-response binding effects between six and twelve seconds after integration and found a duration of RR-binding effects of up to nine seconds. The significance of different durations of binding effects depending on the integrated features for mechanisms in action control is discussed.
4.15pm
Bound to a spider without its web: Task type modulates response executions being influenced by preceding affective information
Lars-Michael Schöpper, Lisann Lötzke, Alicia Jerusalem, & Christian Frings
When responding to a stimulus, action control theories assume response and stimulus features to be integrated into a short episodic memory trace. Repeating any component spurs on retrieval, affecting subsequent response executions. While being reliably observed in non-spatial discrimination tasks, such "binding effects" are typically absent when detecting or localizing a stimulus appearing somewhere; here, usually only a location change benefit emerges, that is, inhibition of return (IOR). Valence has been found to modulate action control processes. In contrast, research on the modulating role of valence on IOR has led to mixed results, with many finding no influence at all. In the current study we tested 1) if task-irrelevant valence is bound to the response and affects subsequent responses due to retrieval, and 2) if the lack of binding and retrieval in localization tasks is responsible for IOR often being unaffected by affective information. Participants discriminated letters (Experiment 1) or localized targets (Experiment 2) on a touchpad in prime-probe sequences. Crucially, during the prime display two images ‐ fruits and spiders ‐ appeared on the touchpad, one of which spatially congruent with the to-be-touched area. In the discrimination task, previously touching a spider compared to a fruit slowed down response repetitions, while the opposite pattern emerged for response changes. In contrast, the localization task only showed IOR. We conclude that 1) task-irrelevant valence is integrated with the response, affecting subsequent responses, and 2) that this is not ubiquitous but depends on task type. The results shed further light on the impact of affective information on actions and offer a possible explanation for an often absent modulation of IOR by valence.
4.30pm
Explicit but no implicit sense of agency for color changes
Julian Gutzeit, Jens Kürten, Lynn Huestegge, & Lisa Weller
The experience to be in control of one's actions and their outcomes is called sense of agency. We wanted to investigate whether subjects feel sense of agency for the color change of a stimulus that they have caused themselves . In two experiments, participants changed the color of certain stimuli by either pressing the corresponding key (study 1) or by fixating the chosen stimulus (study 2). The color change happened after a certain delay contingent to participants' actions (Agency condition). As a baseline, we also implemented stimuli that changed their color continuously and independent of participants' actions after random intervals during the whole trial (Random condition). Participants should experience agency for the color change of stimuli in the Agency condition, but not for the color change in the Random condition. We analyzed explicit agency ratings, as well as interval reproduction errors as a measure of temporal binding. We found higher agency ratings for color-changes when participants believed that they had caused the color-change compared to the condition in which they believed they had no control over the color-change. Surprisingly, we found stronger temporal binding for color changes for which they felt weaker agency compared to color changes for which they felt stronger agency. These findings indicate, in line with previous studies, that explicit and implicit agency measures cannot be used interchangeably and rather suggest that temporal binding might capture other processes than agency or in addition to agency.
Amber Auditorium
Implicit cognition and metacognition
Implicit cognition and metacognition
Session chaired by Raoul Bell
8.30am
Reassessing unconscious priming of numbers: Insights from reaction times, EEG and fMRI
8.45am
Across task adaptation between cognitive and emotional conflict
Elisa Ruth Straub, Moritz Schiltenwolf, Andrea Kiesel, & David Dignath
The cognitive processing of task-irrelevant distractors has potential to interfere with task performance. Our cognitive system has evolved dedicated control mechanisms that suppress interference from task-irrelevant distraction and lead to control-related adjustments in cognitive tasks. However, in real-world situations, we constantly react in the presence of affective stimuli and emotional conflict (such as irony, where emotion portrayed by the face and the voice do not coincide) is an inherent part of human social interactions. But is control confined to ´cognitive´ disturbances or also able to transfer across tasks that induce emotional conflict? We investigate how cognitive and emotional processing interact in Gender Face Stroop Tasks that induce cognitive conflict and Emotional Face Stroop Tasks that were used to study implicit emotion regulation. We hypothesize that conflict-triggered control transfers across the two tasks that comprise the same relevant dimension (i.e., picture of a face) and the same irrelevant dimension (i.e., distractor word). We further asked, how predictability of task-type influences conflict adaptation effects by manipulating the order of tasks in an alternating-two-runs design (Experiment 1, \(n = 89\)) and a quasi-randomized design (Experiment 2, \(n = 88\)). We found reliable across-task adaptation effects from Gender to Emotional Stroop Tasks and vice versa in both Experiments. Our results contribute to observations from neuroimaging studies that demonstrate a common conflict-detection mechanism in Emotional and Gender Stroop Tasks. We propose a domain-general control mechanism involved in monitoring and resolution of conflicts across the cognitive and affective domains.
9.00am
The Implicit Attribute Classification Task (IMPACT) ‐ a new method for investigating automatic appraisals of multidimensional stimuli
Daria Altenburg & Adriaan Spruyt
Past research has demonstrated the important influence automatic appraisals have on a wide range of behaviours, including social interactions, risk taking, consumption choices, and addiction. A widely-used approach in the automatic cognition literature is to measure singular and/or generic appraisals. For example, one of the most commonly used automatic measures, the Implicit Association Test (IAT), assesses one evaluative dimension at a time. However, stimuli typically vary on multiple dimensions simultaneously, raising questions about the appropriateness and ecological validity of this unidimensional approach. To address this issue, we developed the Implicit Attribute Classification Task (IMPACT), a novel implicit measure that allows for the measurement of multiple automatic appraisal dimensions simultaneously and further eliminates both block order effects and the need for target reference categories. Here, we will present results of a study in which we compare the internal and external validity of the IMPACT and the IAT. This work provides a more nuanced approach to measuring automatic appraisals that takes into account the multidimensional nature of even simple stimuli, and has the potential to further our understanding of the influence automatic appraisals have on behaviour.
9.15am
Metacognitive awareness of auditory distraction: Global metacognitive beliefs and stimulus-specific judgements of distraction
Raoul Bell, Laura Mieth, Jan Philipp Röer, & Axel Buchner
The duplex-mechanism account postulates that there are two types of auditory distraction that differ in the metacognitive awareness people have for the distracting effects. Specifically, this two-process account predicts that people should be aware of the disruptive effects of auditory deviants but should have little to no awareness of the interference produced by changing-state in comparison to steady-state auditory distractors. To test this prediction, participants were required to provide metacognitive judgements about the effects of auditory deviant, changing-state, and steady-state distractor sequences on serial recall. In two questionnaire studies, metacognitive beliefs about the effects of the different types of distractor sequences on serial recall were assessed. In another study, participants heard samples of the auditory stimuli before predicting the effects of these stimuli on serial-recall performance. Then, auditory distraction by the different types of distractors was assessed in a serial-recall phase. Finally, participants were asked to give retrospective judgements about the distraction they had experienced. The results suggest that participants are well aware of the effects of changing and deviant stimuli on serial-recall performance. The results thus contradict the idea that there are two types of distraction with differential access to awareness.
Despite many research efforts dedicated towards deciphering the functional architecture underlying metacognition, it is still unclear if there are separable metacognitive systems for different domains of function. In a large sample study (\(N = 155\)) employing pairs of laboratory tasks of metacognition across three experimental domains (visual perception, episodic memory, and attention to action), we examined whether metacognitive ability is determined by universal or modular processes, and whether it is related with various self-report measures of real-world metacognition. Relationships between metacognitive efficiency scores across paradigms and domains were assessed using a combination of correlational and latent variable approaches. Generally, the results point to a mixture of domain-general (unity) and domain-specific (diversity) components. Specifically, Bayesian correlation estimates of metacognitive efficiency as well as confirmatory factor analysis of inter-domain correlations suggested metacognition about perceptual judgments to be mostly domain-specific, whereas convergent indications for interrelations between metacognition in the domains of attention-to-action and memory implied the coexistence of partly specialized metacognitive subsystems. Notably, self-report measures of metacognition were unable to predict metacognitive performance in the laboratory, but represented metacognitive bias rather than efficiency, underscoring prevalent scepticism whether self-report questionnaires provide a useful proxy in metacognition research, as they appear susceptible to potentially unreliable introspections and memory distortions. The results presented in this talk therefore indicate a constitution of both universal and specialized parts for task-based metacognition.
Embodied cognition
Embodied cognition
Session chaired by Sven Hoffmann
10.30am
Valence but not attention allocation moderates the effect of stimulus-hand proximity on conflict processing in a Simon task
Sven Hoffmann, Rico Fischer, & Roman Liepelt
Efficient interaction with the environment requires an adaptation of behavior to the environment. We do so by using cues of our environment and relate these cues to our body to predict outcomes of events. The recent embodied cognition literature suggests that task-relevant stimuli presented near the hands receive more attentional capacity and are processed differently than stimuli, which are presented at a spatially more distant position. Further, it has been assumed that conflict processing benefits from near-hand processing. We tested if this attentional bias in near hand space is modulated by the valence of the to be processed stimulus by combining a cuing paradigm (allocation of visual attention) with a conflict processing paradigm (Simon task) in near vs. far hand space (proximal, distal, respectively). The valence of cue processing was manipulated by using affective (angry vs. neutral smileys) gaze cues. Cue congruency, indicating the correct vs. incorrect attention allocation to the target stimulus onset, did not reveal an effect on Simon compatibility \(\times\) stimulus-hand proximity. In contrast, the interaction of valence \(\times\) Simon compatibility \(\times\) stimulus-hand proximity was significant, as well as the interaction of valence, cue congruency, and hand proximity: For negative valence processing the Simon effect was smaller in the proximal than in the distal stimulus-hand condition. This effect was at least numerically reversed in the neutral valence condition. Regarding the effect of valence on cuing and hand-proximity, the cuing effect was bigger for negative valence in the proximal condition. We conclude that valence, as well as allocation of attention and conflict seem to be a decisive factor for the direction and strength of hand proximity effects.
10.45am
Increased accuracy for stimuli near the hands in a cognitively demanding Simon task
Tillmann Nett, Ruben Ellinghaus, Rico Fischer, Rolf Ulrich, & Roman Liepelt
A prominent finding in embodied cognition is that information is processed differently depending on where it is presented in relation to the body. Thus, several effects are moderated depending on whether the stimuli are presented near the hand or further away from the hands. Specifically, Liepelt and Fischer (2016) found that the Simon effect in a cognitively demanding number categorization task depends on the hand position, with a reduced Simon effect when the stimuli are presented near the hands. In contrast, an increased Simon effect near the hands can be observed in less demanding tasks, such as a color classification task. Liepelt and Fischer (2016) argued that this can be explained by a combination of increased cognitive control and increased processing of spatial features. To identify the mechanism behind the effect more clearly, we re-analyzed the data from Liepelt and Fischer (2016) using Drift Diffusion Modelling and Delta Plots. In our modelling we find a higher desired level of accuracy for stimuli presented near the hands. Furthermore, we found some descriptive evidence that the influence of the relevant stimulus features is increased while the influence of the irrelevant stimulus features is decreased. In addition, the Delta Plots indicate that the time course of the effect is similarly decreasing near the hands and further away from the hands. Overall, these findings are in line with theories predicting that hand-position influences task set configurations prior to stimulus presentation (i,e. proactive rather than retroactive control).
11.00am
Embodied cognition in multitasking: Effects of hand position on crosstalk effects in temporally overlapping tasks
Ruben Ellinghaus, Markus Janczyk, Wilfried Kunde, Rico Fischer, & Roman Liepelt
In contrast to traditional dualistic views of cognition, visual stimulus processing is not independent of bodily factors such as hand positioning. For example, reduced crosstalk between two temporally overlapping tasks has been observed when stimuli are presented near the hands as compared to far from the hands. This result indicates hand-specific attentional processing enhancements which support a more serial rather than parallel processing of the two tasks. To further elucidate how these hand-specific processing modulations affect multitasking performance, the present study consisted of three interrelated crosstalk experiments. Experiment 1 manipulated vertical stimulus-hand distance and stimulus-effect distance orthogonally, with results demonstrating that hand distance rather than effect distance drives the crosstalk reduction. Eexperiment 2 manipulated the physical distance between both hands (i.e., horizontal stimulus-hand & hand-hand nearness), with results showing weakly increased crosstalk when both hands are close to each other. Taken together, Experiments 2 and 1 support the notion of hand-specific attentional processing benefits which can either lead to more serial or more parallel processing of multiple stimuli in view, depending on whether the attentional hand foci of both hands overlap. Experiment 3 is preregistered and currently being run to test this interpretation rigorously. In this experiment, employing an orthogonal manipulation of horizontal nearness and vertical distance, we expect opposing influences of both factors, that is, increased crosstalk for horizontal nearness (replicating Experiment 2) and decreased crosstalk for vertical distance (replicating Experiment 1 and previous work).
11.15am
Hand position in action imagery of typing
Martina Rieger, Stephan F. Dahm, & Victoria K. E. Bart
In action imagery (AI), the body position and therefore actual proprioceptive information about the own body may diverge to different degrees from imagined the imagined body position and therefore imagined proprioception. It is assumed that AI is more difficult with larger discrepancies in actual and imagined body position. In three experiments, we investigated the role of the hand position during imagined typing and analyzed typing durations and attentional focus. In all experiments, participants performed an execution condition, an imagination condition in which the hands were lying flat next to the right and left side keyboard, and a second imagination condition. In the second imagination condition, the hands were either next to the keyboard with clenched fists (Experiment 1), lying flat on the right and left side of the screen (Experiment 2), or on the back of the participant (Experiment 3). Concerning durations, no significant differences between conditions were observed in Experiment 1. However, in Experiment 2, imagination with hands on the sides of screen was significantly shorter than execution, and in Experiment 3, imagination with hands on participant's back was significantly longer than imagination with hands next to the keyboard. In all experiments, systematic differences between conditions were observed in attentional focus. For instance, participants paid more attention to the movement of fingers in most imagination conditions than in execution conditions. In conclusion, the data provide evidence that incongruence between proprioception/ actual body position and imagined proprioception/ imagined body position changes the content of AI.
11.30am
Visual perspective taking in older adults: A HD-tDCS study
Mandy Roheger, Anna Mäder, Andrew K. Martin, & Marcus Meinzer
Background: Aging can result in impaired socio-cognitive processing, such as visual perspective taking (VPT). A recent transcranial direct current stimulation (tDCS) study demonstrated that the right temporoparietal junction (rTPJ) is causally and selectively involved in VPT in young adults. Here, we investigated if the same relationship exists in advanced age. Methods: 60 healthy older adults were stratified to receive either rTPJ or dorsomedial prefrontal cortex (dmPFC) focalized anodal tDCS in a sham-controlled, double-blind, repeated-measures design while performing a VPT task with two conditions. During visual perspective-tracking (level 1 VPT), participants were asked whether an avatar can see a visual stimulus; during visual perspective-taking (level 2 VPT), participants were asked whether the stimulus was presented on the left or right side of the avatar's position (requiring embodied rotation). Rotation angle of the avatar and body position of the participants in front of the screen varied. 30 young participants completed the VPT task without stimulation. Results: Both older and younger adults show an effect of rotation angle (slower reaction times with greater angle) and body posture (lower reaction times for incongruent body posture with avatar) during the perspective taking task. Older adults were significantly slower in responding, across tasks and stimulation conditions. rTPJ tDCS did not affect embodied perspective taking in older adults. DmPFC stimulation resulted in further slowing of response latencies in older adults. Discussion: This study failed to replicate modulatory effects of focalized rTPJ-tDCS. Potential mediators of stimulation outcome (e.g. neural network organization, electrical current flow) are currently being explored.
11.45am
The cognitive penetrability of representational momentum effects by unrelated factors
Jannis C. Friedrich, Laura Voigt, & Markus Raab
Our cognition represents the physical momentum of an object, called representational momentum. When individuals are shown a moving target which is then occluded, they tend to misremember it as further along its trajectory, with a distance between the true and remembered position (this distance being called displacement). It has been demonstrated that this effect is cognitively penetrable, changing when manipulating factors such as weight or velocity of the target. Yet, so far studies have focused solely on manipulating factors also related to momentum in our physical world. The current study tests the hypothesis that representational momentum can be altered by manipulating qualities unrelated to physical momentum; for instance, recent success of the target. We hypothesise that a target perceived as 'successful' will result in larger displacement, because success is conceptualised as 'progress' and 'forward impulse'. In an online study using a randomized within-subjects design, 120 participants (60 per condition, from online participant pool Prolific) watch a 3-second video of a football player either scoring (successful, 12 trials) or missing a penalty kick (unsuccessful, 12 trials), followed by a 4-second video of this player running. This video is then occluded and participants are asked to indicate the final position of the player by moving a slider to what they remember as the final frame of the video. The outcome measure is the number of frames between true and remembered frame. Significance of this research are theoretical implications for visual cognition research on invariant representations, as well as support for a notion from ecological cognitive approaches, that environmental constraints form our cognition.
The role of modal and amodal representations in cognitive functions
SymposiumThe role of modal and amodal representations in cognitive functions
Symposium organized by Karin Maria Bausenhart and Nicoletta Simi
Historically, it was often assumed that central cognitive functions, such as thinking, decision making, and speech comprehension, are based on abstract or symbolic cognitive representations. Such representations would not directly preserve the structural properties of their physical referents, but encode them in an amodal fashion, for example, in the form of propositional or schematic representations. Therefore, sensory-motor experience based on specific modal properties of the referents should have little impact on central processing. In contrast, alternative theoretical approaches proceed from a more embodied perspective. Such accounts suppose that cognitive processes operate on rather concrete and experiential representations. Accordingly, such representations would still preserve the modal properties of their referents, and these modal properties may in turn impact on the results of cognitive operations. There is increasing evidence that such modal representations based on sensory-motor experience not only emerge in perceptual and motor stages of processing, but also play an important role in more central cognitive processes. To reconcile these originally opposing views, it seems promising to assume that cognitive representations at all processing stages may range on a continuum from modal to amodal, where different representational formats may inform various cognitive operations and serve different goals. In this symposium, we will explore the role of modality, embodied experience, and abstraction in various aspects of cognitive processing, such as language production (Abdel Rahman & Vogt), cognitive control (Simi et al.), multitasking performance (Koch & Stephan), orientation behavior (LeVinh et al.), and motor planning (Bhatia et al.).
4.00pm
Embodied language production: Is speaking shaped by experientially grounded meaning?
Rasha Abdel Rahman & Anne Vogt
How is language production shaped by our sensory-motor experiences? During verbal communication our sensations and (bodily) experiences are frequent topics of conversation. However, while some meaning relations as categorical, associative or thematic links are known to form the basis for lexical-semantic processing during language production, little is known about whether and how lexical selection during speech planning is shaped by sensorimotor experiences. I will present experiments on the role of experientially grounded meaning during lexical-semantic processing. The results suggest that experiential traces form one aspect of the many facets of meaning that determine which words we select to express an intended meaning. I discuss the relation between experience-related meaning and traditionally investigated semantic factors in language production research and conclude with a discussion of the difficult distinction between "semantic" and "embodied" manifestations of sensorimotor experiences in language production.
4.15pm
Cognitive control mechanisms in a reading task involving semantic ambiguity
Nicoletta Simi, Ian Grant Mackenzie, Hartmut Leuthold, Markus Janczyk, & Carolin Dudschig
Cognitive conflict is seen as a crucial factor in triggering subsequent adjustments in cognitive control. Studies have suggested that the implementation of control following conflict detection might be domain-general in that conflict experienced in the language domain recruits control processes that deal with conflict experienced in non-linguistic domains. During language comprehension, humans often must recover from conflicting interpretations as quickly and as accurately as possible. In the present study, we investigate how people adapt to conflict experienced during processing semantically ambiguous sentences. Experiments 1‐3 investigated whether semantic conflict produces CSEs within a subsequent Stroop task and whether Stroop conflict leads to adjustments in semantic processing. Experiments 4‐6 investigated whether such semantic conflict results in conflict adaptation in subsequent sentence processing as reflected by the congruency sequence effect (CSE). Although processing conflict was consistently experienced during sentence reading and in the Stroop task, and in control blocks conflict adaptation was observed within the manual Stroop task, we did not observe any within-task or cross-task adaptation effects. Specifically, there were no cross-task CSEs from the linguistic task to the Stroop task and vice versa (Experiment 1‐3) ‐ speaking against the assumption of domain-general control mechanisms. Moreover, experiencing conflict within a semantically ambiguous sentence did not ease the processing of a subsequent ambiguous sentence (Experiment 4‐6). We argue that the present zero control on language processing must be interpreted with caution since the timing of when conflicting information is presented and the type of conflict might be critical.
Multitasking refers to the requirement to coordinate the processing of two tasks that are performed either sequentially (task switching) or simultaneously (dual tasks). Relative to performing a single task, multitasking usually incurs performance costs. Standard theoretical approaches assume that these are due to amodal "central" structural limitations, such as a time-consuming "task-set reconfiguration" process (in task switching) or a structural, serial all-or-none decision and response-selection "bottleneck" (in dual tasks). The notion of a "central" process thus implies independency of sensory or motor processes, thus implying amodality. However, while the notion of modality-specific processing resources has some tradition in dual-task research (e.g., suggested by Wickens, 1984), there is little theorizing about the role of modality in work on sequential task switching, presumably because this experimental paradigm has been used to examine cognitive control, which is defined in rather abstract terms. In the presentation, we describe empirical effects showing modality-specific influence on the task mappings (e.g., auditory-vocal tasks, visual-manual tasks) in task switching, and we propose a theoretical account in terms of action control mechanisms based on online prediction of intended action effects ("ideomotor" account). Specifically, we propose that modality-specific confusion of separable task mappings is a major mechanism underlying multitasking "crosstalk" in action selection.
4.45pm
Place-name categorisations affect wayfinding behaviour
Lilian LeVinh, Jule Wildt, & Hanspeter A. Mallot
Regionalisation is the grouping of places into regions. It has been shown to affect spatial reasoning in path planning: when choosing the shortest route to a goal where two equidistant routes exist, humans prefer the route with less interregional border crossings (Region dependent route preference, RDRP). Perceived regionalization depends on the semantic and contextual similarity of landmark objects, but may also be induced by place names. However, in this case only by contextual similarity. Here we address the role of place name semantic similarity in wayfinding in a ring-shaped maze with 12 places, numbered 1-6 in the ring, each with an attached dead end. Places were marked by object names allowing two possible categorizations resulting in different regions (either places 1-2, 3-4, 5-6 or 2-3, 4-5, 6-1, each with their attached dead end). Prior to the wayfinding task, subjects were asked to categorize the landmark names into three groups of four, and thereby discovered one or the other categorization. We tested whether subjects show RDRP and whether RDRP changes depending on the previously discovered categorization. In the test phase subjects were asked to take the shortest route between three places. Half of the tasks could be solved by taking either of two equidistant routes. The routes differed however in that depending on the previous categorisation one route will lead to more border crossings between regions than the other. We found that subjects show RDRP according to their previous categorization. Therefore, place-names grouped into semantic categories can evoke a significant regionalisation. Also, the regionalisation effect can lead to different orientation behaviours even within the same environment.
5.00pm
Reviewing evidence for different representations in perception and action
Kriti Bhatia, Markus Janczyk, & Volker H. Franz
The Perception-Action Model (PAM) assumes that visual information in the human brain is processed in qualitatively different ways by dorsal and ventral cortical streams. Evidence for differential processing in the two streams comes from the Garner paradigm, that tests selective attention to stimulus features. It has a baseline condition, where only task-relevant stimulus variation occurs, and a filtering condition, where task-relevant and irrelevant variation occurs. Longer reaction times in the filtering condition compared to the baseline condition indicate Garner interference. Ganel & Goodale (2003, Nature) observed Garner interference in perception and manual estimation (both arguably based on ventral/amodal representations), but not in a visuo-motor action task (arguably based on dorsal/modal representations). We replicated this experiment (\(N = 24\)) in a within-subjects design. In the perception task, participants pressed a button to judge rectangular objects as "narrow" or "wide". In the manual estimation task, they estimated the width of these objects with their finger-span. In the action task, they grasped these objects along the width. We confirmed previous findings in perception and grasping. However, we found no Garner Interference in manual estimation start time. Frequentist and Bayesian analyses of our results and a review of other, similar studies reveal inconsistencies in the literature. Hesse & Schenk (2013, Behavioural Brain Research) proposed an alternative account for Garner interference that explains our results in manual estimation, but casts doubt on Garner interference as valid evidence for the PAM. We extend this framework and assess claims of different processing and representations in the two streams assumed by the PAM.
Sapphire Court
Attention I
Attention I
Session chaired by Jochen Laubrock
8.30am
Occipital transcranial magnetic stimulation (TMS) reduces presaccadic attention
8.45am
Does temporal preparation influence bottom-up salience in spatial selection? Evidence from the N2pc
Janina Balke, Bettina Rolke, & Verena Carola Seibold
Previous research has established that temporal preparation facilitates spatial selection in visual search. In two ERP studies, we aimed to clarify how specifically temporal preparation influences spatial selection. According to visual search models, spatial selection of stimuli is affected by bottom-up salience, top-down goals, and selection history. We investigated whether temporal preparation influences bottom-up salience. To this end, we employed a visual search task in which participants (\(N = 24\) in both studies) searched for a salient target among homogenous distractors, and we measured the N2pc as an index of spatial selection. To manipulate bottom-up salience, we either presented a salient but task-irrelevant distractor (Study 1), or we varied the salience of the target (Study 2). To manipulate temporal preparation, we presented a warning signal before the search display and varied the interval between warning signal and search display (blocked foreperiod paradigm). In both studies, temporal preparation accelerated spatial selection of the target, as indexed by a reduction in the onset latency of the target N2pc. In Study 1, temporal preparation additionally modulated the amplitude of the N2pc elicited by the salient but task-irrelevant distractor. Furthermore, in Study 2, temporal preparation reduced the latency of the target N2pc to a stronger extent for low salient targets than for highly salient targets. Together, the results of the two studies provide evidence that temporal preparation facilitates spatial selection by influencing bottom-up salience.
9.00am
Combined visual-auditory fear-related stimulation delays the disengagement of spatial attention from an invalidly cued position: An ERP study
Marlene Pacharra, Ulrike Zimmerpresenting, & Mike Wendt
Spatial shifts of attention evoked by a peripherally presented visual or auditory cue can be enhanced by adding fear-related content, such as the picture of a face displaying an anxious expression or a fearful vocal sound, compared with emotionally neutral cue stimuli (Zhang et al., 2017; Zimmer et al., 2016). We investigated this "emotional enhancement" of attentional cueing by presenting redundant visual-auditory cues (i.e., a neutral or fearful face together with a neutral or fearful voice, located on the same side) which preceded an (emotionally neutral, visual) target stimulus, displayed on the same side (i.e., valid cueing condition) or on the opposite side (i.e., invalid cueing condition) and recorded attention-related ERPs (i.e., P1, P3). In reaction times, a typical cueing validity effect was only observed when both cues were related to fear. Corresponding with this behavioral effect, the P1 displayed a maximum difference between valid and invalid trials in the "both cues fearful" condition. A main effect of validity in the P3a component (i.e., larger P3a in invalidly than in validly cued trials) suggested, however, that spatial attention was first drawn to the cued position for all types of cue combinations. The subsequent P3b displayed a pattern consistent with the notion that redirecting of attention in invalidly cued trials was impaired when both cues were associated with fear. In conclusion, our results suggest that bimodally presented fear-related cues lead to enhanced focusing of attention making it more difficult to disengage attention from the cued location.
9.15am
Decoding non-spatial attention sets in PFC and the functional/anatomical connectivity fingerprints of attention-related control structures in PFC
Daniel Baldauf
Prefrontal cortex contains two important areas, the frontal eye field (FEF) and the inferior frontal junction (IFJ), that are both involved in the orchestrating functions of attention, working memory and cognitive control. Here we studied both the representational content within PFC (based on MEG-MVPA decoding) as well as its functional and anatomical connectivity patterns (based on temporally high-resolving rsMEG and DTI, respectively). Multivariate pattern analysis in auditory and visual tasks revealed the selectivity of the inferior frontal junction to non-spatial information, e.g., feature- or object-based information, allowing for the successful decoding of non-spatial attention, independently of task modality. In terms of PFC connectivity, we found striking differences in how these regions are connected to other cortical networks: while FEF is more strongly connected with topographically organized areas, IFJ shows systematically stronger connection with sensory areas that lack such a clear topographic organization (predominantly in the ventral visual stream). As both the representational content and the connectivity fingerprints are a crucial pre-requisite for structures in PFC to provide top-down attentional control signals, our data points at a systematic division of labor between FEF and IFj in the control of spatial versus non-spatial attention, respectively.
9.30am
Please ignore this: Early visual cortex and precuneus activation as neural correlates of distractor preparation in negative cueing
Antonia Jüllig, Kenji Yamauchi, Jun Kawahara, & Stefan Pollmann
Selecting relevant from irrelevant input enables us to act in a sensory rich environment. This attentional filter includes target facilitation and distractor rejection. Distractor foreknowledge can facilitate visual search, yet the neural mechanism is unclear. Does working memory store distractor features as "templates for rejection" biasing attention away from them (Woodman & Luck, 2007), or are they attended before being rejected, according to the "search and destroy" theory (Moher & Egeth, 2012)? As most negative cueing studies used color stimuli, we also asked if their findings generalize to shapes as feature conjunctions. Two samples of twenty-five healthy humans participated in two fMRI studies, with the first one presenting a distractor shape cue, a target cue, and a task-irrelevant control condition in a randomized order, and the second one presenting the conditions in a blocked way to control for top-down inhibition due to task-set changes. Both experiments showed early visual cortex pre-activation with distractor- compared to task-irrelevant cues, in line with spatial and implicit distractor preparation paradigms. Precuneus activation distinguished distractor- from target preparation, against the "search and destroy" assumption that distractors are represented like targets before search. We confirmed a negative cueing effect for shapes, yet its neural signature differed from that for colors (Reeder, Olivers, & Pollmann, 2017), suggesting that the underlying mechanism depends on stimulus complexity.
9.45am
Multimodal cohesion affects attention and interpretation: An eye-tracking study of film
Jochen Laubrock
Multimodal cohesion is a method that describes and analyzes the means by which narrative discourse structures are "bound together" by information from different domains, e.g., visual, auditory, and verbal. Here we present an exploratory experimental investigation of how multimedial cohesion guides visual attention and narrative understanding. Using a cohesion analysis we identified important multimodal cues that film directors use to establish the setting of a scene and its semantic interpretation. We manipulated film scenes by blurring important cues to the setting or altering the soundtrack or verbal cues, and recorded viewers' eye movements as well as questionnaire data about the scene. Eye movement recordings suggest that removal of cues (a) caused less attention to be allocated to the blurred parts of an edited scene—which can be considered as a manipulation check, and (b) led to uncertainty about the setting of the subsequent unedited scene, which triggered active search behaviour in order to reduce that uncertainty. Questionnaire results supported this interpretation: Viewers of modified scenes were less certain about the specific identity of the setting, and replaced it with a more generic identity. Taken together, our results using experimental manipulations of film material show that film directors employ multimodal cohesive cues to guide their viewers attention and narrative comprehension.
Attention II
Attention II
Session chaired by Jochen Laubrock
10.30am
How to extend the attentional blink: Effects of working memory updating and load
Jochen Laubrock
The attentional blink is a phenomenon of temporal attention, showing that we can temporarily be blind for stimuli following a task-relevant stimulus. We and others have shown that working memory encoding is critical in eliciting the attentional blink. Here we investigate how working memory load and working memory updating affect the attentional blink. In Experiment 1 the task was to report the first and last letter of an RSVP stream. Targets T1 and T2 were embedded either in a stream of digits, or in a stream that began and ended with digits, but temporarily switched to letters between T1 and T2, requiring permanent working memory updating in order to store the last letter. The digit stream produced the classic attentional blink, whereas the switched stream led to an extended and strong attentional blink that never recovered. Thus, working memory encoding and updating are critical components of the attentional blink. In Experiment 2 we studied the effects of working memory load on the attentional blink. Subjects had to identify two letters in a stream of digits, and store the last distractor digit for later report. Thus we effectively combined the attentional blink and the OSPAN task, using memory loads of 0 to 4 items. Target detection performance was reduced with higher memory loads, but the effect was purely additive and did not affect the attentional blink. Taken together, working memory encoding and updating, but not working memory load affect the attentional blink.
10.45am
Self-generated vs. instructed strategies ‐ strategy use in a change blindness study
11.00am
Oculomotor evidence for inter-individual differences in the acquisition and adaptation of long-term context memories: The role of global processing bias and executive control
Artyom Zinchenko, Markus Conci, Hermann Müller, & Thomas Geyer
Visual search becomes more efficient if the target item is consistently encountered within a stable spatial arrangement of distractors, as such context is thought to guide visual search (contextual cueing; Chun & Jiang, 1998). The default local relative to global processing bias in participants was shown to further facilitate context learning (Bellaera et al., 2014). In contrast, changes of the target location in a nevertheless constant array weakens contextual cueing. Here, we tested whether global-local bias dichotomy, as well as participants' executive control capacities and mental rotation abilities can modulate the acquisition and subsequent adaptation of already formed context memories. We also recorded participants' eye-movements to explore whether the potential modulations would be expressed both in the reaction times and oculomotor measurements (number of fixation, saccade amplitudes, exploratory and confirmatory saccades). As a result, the cueing effect was reduced significantly in the relocation phase after it was established successfully in the initial learning phase. Additionally, while the default local bias indeed facilitated initial acquisition of context-memories (in line with Bellaera et al., 2014), this effect was reversed in the relocation phase, where more global attentional bias, as well as enhanced executive control, resulted in more efficient adaptation in both behavioral and eye-movement measurements.Together, these findings provide behavioral and oculomotor evidence for the role of inter-individual differences in the initial acquisition and subsequent adaptation of long-term context memories. Context learning and adaptation are not only dependent on physical properties of stimuli, but also on observers' attentional bias and executive control.
Warning signals temporarily increase arousal, which has been attributed to performance enhancements. This phasic alerting effect occurs among other sensory modalities with auditory and visual cues. A prominent hypothesis is that auditory warning signals induce greater beneficial effects than visual ones. However, it is unclear if indeed audition dominates over vision or if it is mediated by the stimulus intensity of the warning signals. Here, we parametrically assessed the effects of visual and auditory alerting cues with matched stimulus intensities based on the known reaction times of a simple target detection task. Participants conducted a speeded choice-reaction task with visual, auditory and no cues. Results showed that visual and auditory warning signals generally boosted reaction times. However, we found no evidence for greater phasic alerting effects in auditory warning signals over visual warning signals. Thus, these findings suggest that auditory and visual alerting can be equally effective when the intensities of the stimuli are matched.
11.30am
How does attention influence the effects of color categories
Aimee Martin & Karl R. Gegenfurtner
Color is a prime example of categorical perception. Previous studies have shown that color discrimination is usually faster and more accurate across color categories (e.g., blue vs green), compared to discrimination of colors from the same category (e.g., shades of blue). Whether attention may be driving this perceptual categorical effect is unclear. Within attention, feature based accounts have demonstrated that only colors that match a pre-defined target capture attention, however few studies have investigated the role of color categories. We investigated whether these attention effects depend on color categories, or whether they are driven by color similarity alone (i.e., distance in color space). In a spatial cueing task, participants searched for either a blue or green singleton target interleaved across blocks. Colors were presented in DKL color space. Prior to the target display, cues unpredictive of the target location were briefly shown, this contained a singleton cue that varied in different hues of blue to green. At the end of the experiment, participants were shown color patches of the hues used and asked to rate the hues used as 'green' or 'blue' and to give a rating on their confidence from 1 to 5 (i.e., 1 = guessing, 5 = highly confident) to measure the blue-green border of the hues used for this sample. Our results showed validity effects (e.g., RTs were quicker on valid trials compared to invalid trials) only when cues were from the same category as the target (e.g., blue cues for the blue target; green cues for the green target), with no observed validity effects once cue color crossed into the next category. Overall, these results suggest that attention contributes to categorical effects for blue and green colors.
11.45am
Effects of spatial influences on inhibition of return in online studies
Paula Soballa, Lars-Michael Schöpper, Christian Frings, & Simon Merz
Inhibition of return (IOR) refers to the phenomenon that reaction times are slower to a target when it is presented at the same location a cue was shown shortly before, compared to conditions with previously uncued locations. This response delay is often preceded by a short period of the reversed effect, the facilitation of return (FOR), where a location repetition between cue and target leads to faster responses. IOR has been investigated for nearly 40 years with a lot of different stimuli, in variants of the standard design, and across modalities. Yet, it is still unclear, whether spatial biases influences IOR as stated by Spalek and Hammad (2004) or if the observed differences are better explained through general reaction time differences depending on the target position (Snyder & Schmidt, 2014). Further, neither IOR nor FOR have been systematically investigated online yet. With the aim to clarify both questions, three online studies (each \(N = 31\)) were conducted and different time intervals and target locations were systematically examined. IOR was reliably observed with cue durations of 50 and 300ms as well as different interstimulus intervals ranging from 0 - 900ms. Furthermore with shorter stimulus onset asynchronies, FOR was observed. Essentially, IOR effects were moderated by target locations and the main effect of target position even revealed a lack of IOR at the lower location in all three experiments. Crucially, those results cannot be explained as a function of general reaction time differences and target position.
Advances in TVA-based visual attention research: Progress toward new stimuli, tasks, and data
SymposiumAdvances in TVA-based visual attention research: Progress toward new stimuli, tasks, and data
Symposium organized by Jan Tünnermann and Ingrid Scharlau
The Theory of Visual Attention (TVA) understands visual perception as a competition in which visual categorizations compete for encoding into visual short-term memory, biased by attentional and perceptual parameters. With its mathematical framework, TVA formalizes this process rigorously, linking data to theoretical concepts. A trend that has now lasted several years is that some TVA research is moving toward new and more complex stimuli, tasks, and data. Continuing our last-year symposium on this development, we discuss work on how TVA can be applied to a broad range of experimental data and situations: Banh and Scharlau report effects of negation in instructions on visual attention and show how TVA can help to understand the processing of negation. While TVA has already been applied to study the effects of visual salience on attentional processing in the past, social salience effects have not been studied with TVA so far. Scheller et al. present recent work that studies such effects (salience induced by self-association) with TVA. The influence of eye and hand movements on TVA parameters is in the focus of Kreyenmeier et al.'s work. Kristjánsson and Tagu investigate the benefits of the visual-foraging tasks for assessing the flexibility of attentional control in continuous behavioral situations. Tünnermann and Schubö connect to this with a study of the dynamics of attentional-template switching in visual foraging with the help of TVA-based simulations. A novel extension of TVA for including reaction time data in TVA analyses is presented by Blurton et al.
4.00pm
Effects of verbal negation on TVA's capacity and weight parameters
Ngoc Chi Banh & Ingrid Scharlau
How does verbal negation influence attention? Although rarely asked in research, it stands to reason that instructions using a negation might hamper the distribution of attention. The present work investigates the influence of verbal negation on attention by combining the theoretical concepts of pragmatic frames and TVA. The concept of pragmatic frames allows formalizing the interaction between two partners intending to attain a joint goal. According to Rohlfing et al. (2016, Frontiers in Psychology), pragmatic frames can be defined as sequences of collaborative actions performed with a partner, involving operations on the cognitive as well as communicative and pragmatic level. Interaction history is a key aspect in pragmatic frames. Repeating a particular situation leads to familiarization and enables partners to develop expectations regarding how actions relate to each other. The present work examines the cognitive aspect of interaction history. Attentional parameters were assessed via unspeeded temporal-order judgement tasks. Study results indicate that—depending on the setting—the overall available processing capacity can be affected by negation. We will discuss repetition effects on visual capacity between the negation and affirmative condition and relate our findings with other studies assessing TVA parameters in dual-task situations.
4.15pm
Using TVA-TOJ to probe effects of social salience on early attentional selection
Meike Scheller, Jan Tünnermann, Katja Fredriksson, & Jie Sui
The way in which we respond to information in our environment is strongly influenced by the perceptual salience it exhibits. Furthermore, the degree to which information is related to ourselves, compared to other individuals, alters response performance in a way that mimics the effects of perceptual salience. A large body of evidence from the last decade now demonstrates how mere association of the self and other social identities with arbitrary sensory information can trigger a cascade of changes in perception, memory, decision-making as well as motor control, leading to behaviourally meaningful and robust processing benefits of self-related information. The present study makes use of the flexibility of the TVA-informed temporal order judgement task to assess how early attentional selection sets up self-prioritization by altering attentional weights or processing rates with which socially loaded information is encoded into short-term memory. That is, self-prioritization may be mediated either by an increase in attentional weights towards the self-associated stimulus, or by an additional change in processing capacity that speeds up processing of self-related information or slows down processing of other-related information. Preliminary results (\(N = 70\)) suggest that the tendency to prioritize self-associated information during attentional selection is rather reflected in a shift in attentional weights towards the socially relevant stimulus, while changes in processing capacity following social association seem negligible. This tentatively suggests that the relative social salience of self- over other-related information triggers a selective reallocation of attentional resources between the socially-loaded perceptual objects, without increasing the system's capacity.
4.30pm
Using the theory of visual attention (TVA) to measure premotor attention
Philipp Kreyenmeier, Heiner Deubel, & Nina M. Hanning
Eye and hand movements are preceded by shifts of visuospatial attention towards the motor targets. Whether attention is allocated independently towards multiple effector targets during combined eye-hand movements, or whether a single attentional system underlies target selection for multiple effectors, has been debated. Independent target selection mechanisms would predict an increase of overall attention capacity with the number of active effectors, whereas a single selection mechanism would predict a fixed attention capacity and competition between effectors. Here, we directly test these two competing hypotheses using the Theory of Visual Attention (TVA; Bundesen, 1990) to measure premotor shifts of visuospatial attention. In two experiments, observers (\(n = 8\)) performed eye, hand, or combined eye-hand movements to centrally-cued locations and reported letters briefly presented before the execution of a saccade or reach movement. Modeling the data according to TVA revealed that attention was predominantly allocated to the movement-relevant locations‐without competition between effectors. However, the parallel benefits at the saccade and reach targets had concomitant costs at movement-irrelevant locations, resulting in a fixed overall attention capacity. Moreover, we show that premotor shifts of attention dominate over voluntary deployment of processing resources, yielding severe impairments of voluntary attention allocation. Our findings suggest that attention shifts to multiple effector targets without mutual competition, as long as sufficient processing resources can be withdrawn from movement-irrelevant locations. They also demonstrate that the TVA framework can be used as a sensitive tool to measure action-related shifts of visual attention.
4.45pm
On the flexibility of attentional control in continuous behavioral tasks
Árni Kristjánsson & Jerome Tagu
The theory of visual attention (TVA) has been considered by many to be the best candidate of current attentional accounts to explain patterns of human performance during continuous selection tasks where observers need to select a multitude of different targets among distractors (typically referred to as foraging tasks). The theory has the flexibility required to account for factors that can influence performance, such as value, salience and discriminability, and proximity. I will present new data from foraging by eye gaze and mouse selection. In this paradigm the value of different target types varies, to address how different priority settings affect foraging performance. Our observers performed four foraging tasks differing by selection modality and target value. During gaze foraging, participants had to accurately fixate the targets to select them and could not anticipate the next selection with their eyes, while during mouse foraging they selected the targets with mouse clicks and were free to move their eyes. We manipulated both target value and proximity. We found notable individual differences in search strategy, confirming the existence of internal biases towards value, proximity and priming. Initial steps at modelling these findings within a TVA setting will be presented.
5.00pm
Integrating attentional template switching dynamics in visual foraging with TVA
Jan Tünnermann & Anna Schubö
In recent years, visual foraging tasks in which participants search and collect multiple target types from crowded displays (or "patches") have been used to assess the flexibility with which attentional templates are switched during the search. The feature context of a patch modulates this flexibility. For instance, color-defined targets lead to efficient search and frequent switching, while conjunction targets (defined by combining features) lead to slower foraging and extended same-type selection runs. In earlier work, we developed TVA-based simulations showing how TVA's attentional control mechanics predict fewer visual short-term memory operations (and possibly less effort) when employing run-like selections in conjunction foraging. These simulations included static rules for switching attentional templates (e.g., "if no more targets of type 1 are found, switch to type 2"). Here we extend this work by modeling more realistic implicit template switching dynamics based on inter-target priming and template memory decay. We compare predicted run statistics and parameters of foraging efficiency with data from visual foraging experiments with humans. Moreover, we discuss how TVA can be extended toward more naturalistic paradigms such as visual foraging by modeling the dynamic interactions between environment, attentional control, and visual short-term memory content.
5.15pm
A TVA-based account of response times: Current developments and future directions
Steven Blurton, Søren Kyllingsbæk, Anders Petersen, Claus Bundesen, & Signe Allerup Vangkilde
In this talk, we present recent theoretical advancements in TVA that include a TVA-based account for response time (RT) distributions and outline future directions that are based on these advancements. Although the scope of TVA is more general, TVA-based research so far has focused on accuracy-based tasks using highly discriminable stimuli to investigate mechanisms of attention on selection of categories (pigeonholing) and selection of elements (filtering). We have recently extended this framework to describe perceptual categorisation in mutually confusable stimuli. The account is compatible with many other theories, serving as a back-end to explain perceptual decisions depending on exponential processing rates. Using TVA as a front-end, we have tested the new model with data from single stimulus recognition tasks with two or more perceptual categories. We present new applications to speeded responses to targets presented among distractors. The new model can account well for feature search distributions just by incorporating the filtering mechanism of TVA. The model is also applicable to conjunction search, but this required introducing of a novel set of weights that describe possible serial processing, effectively combining TVA with Guided Search. Newer directions include applications of the model in tasks that involve cognitive control, such as the Simon task. For this, we added a TVA-like response selection stage to the model and use the theory of Executive Control of TVA (Logan & Gordon, 2001, Psychological Review) to make the model appliable to data obtained in experiments that include response conflicts.
Sunshine Saloon
Face perception from laboratory to naturalistic contexts
SymposiumFace perception from laboratory to naturalistic contexts
Symposium organized by Mario Reutter and Matthias Gamer
Humans are visual creatures. While other animals may recognize conspecifics by olfactory or auditory signals, humans mostly employ visual cues originating from the face. Having relied on this information for millennia, Homo sapiens has developed an outstanding capability to recognize and interpret faces under varying conditions. This ability has been studied for decades using still images, but recent technical developments provided novel approaches that are more immersive and interactive. In this symposium, different perspectives on face perception are highlighted, ranging from controlled laboratory settings to ecologically valid naturalistic contexts. First, we learn how 3D scanning of real faces enables a dissociation of texture and shape information in face recognition. Second, the impact of attentional deployment towards distinct facial features are discussed with respect to the phenomenon of fear generalization. Third, we look at reliable differences in social attention across visually rich images and their relation to social anxiety or alexithymia. Fourth, we shift focus to the interplay between self-focused attention and social anxiety during live video calls. Finally, gaze coordination in a real social interaction between parents and children is addressed by presenting data from a minimally invasive dual eye-tracking setup. Taken together, this symposium covers the broad spectrum from isolated face perception to social information processing in complex scenes and it integrates findings on face recognition, social attention, and mechanisms of social interaction. This multi-angled approach, closely linked to the conceptual framework of cognitive ethology, is crucial to both accurately describe social phenomena and to identify the underlying processes.
8.30am
Contributions of shape and texture to learning and recognition of same and other-"race" faces
Jürgen M. Kaufmann, Xiaomei Zhou, Marlena Itz, Sandro Vogt, Stefan R. Schweinberger, & Catherine Mondloch
It has been suggested that texture information contributes more to familiar face recognition than shape and that there is a shift regarding the relative importance from shape to texture with increasing face familiarity. Building on previous research and using stimuli derived from a 3D camera system, we tested contributions of shape or texture information in a learning experiment. Major aims were i) to replicate findings of a dominant role of texture for face recognition and ii) to test whether these findings also extend to other-"race" faces. Note that although the other-"race" effect (ORE; superior recognition of own- relative to other-"race" faces) is well established, the mechanisms underlying it are not well understood. In the present behavioural study, we compared the influence of shape and texture information on Caucasian participants' learning and recognition of Caucasian and East Asian faces using two complementary approaches. In Experiment 1, participants studied veridical, shape-caricatured, or texture-caricatured faces and then were asked to recognize them in an old/new recognition task. In Experiment 2, all study faces were "veridical" (i.e. not caricatured), but we independently removed the diagnosticity of shape (or texture) cues in the test phase by replacing a face's original shape (or texture) with average face shape (or texture). Despite an expected overall own-"race" advantage, participants' use of shape and texture cues was comparable for own- and other-"race" faces, replicating previous findings that argue for a more dominant role of texture for recognizing learned faces. In addition, the results suggest that the other-"race" effect is not attributable to qualitative differences in the use of shape and texture cues.
8.45am
Individual patterns of attentional exploration predict the extent of fear generalization in humans
Mario Reutter & Matthias Gamer
Generalization of fear is considered an important mechanism contributing to the etiology and maintenance of anxiety disorders. Although previous studies have identified the importance of stimulus discrimination for fear generalization, it is still unclear to what degree overt attention to relevant stimulus features might mediate its magnitude. To test the prediction that visual preferences for distinguishing stimulus aspects are associated with reduced fear generalization, we developed a set of facial stimuli that was meticulously manipulated such that pairs of faces could either be distinguished by looking into the eyes or into the region around mouth and nose, respectively. These pairs were then employed as CS+ and CS− in a differential fear conditioning paradigm followed by a generalization test with morphs in steps of 20%. Shock expectancy ratings indicated a moderately curved fear generalization gradient that is typical for healthy samples but its shape was altered depending on individual attentional deployment: Subjects who dwelled on the distinguishing stimulus regions faster and for longer periods of time exhibited less fear generalization. Although both, pupil and heart rate responses also showed a generalization gradient with pupil diameter and heart rate deceleration increasing as a function of threat, these responses were not significantly related to patterns of visual exploration. In total, the current results indicate that the extent of explicit fear generalization depends on individual patterns of attentional deployment. Future studies evaluating the efficacy of perceptual trainings that aim to augment 18 stimulus discriminability in order to reduce (over-)generalization seem desirable. Repository: https://osf.io/4gz7f/.
Humans show a robust tendency to attend to faces in images, but also differ consistently in the strength of this attentional preference. Previous research remained inconclusive as to how a stronger attentional preference for faces may be indicative of an individual's personality or clinical characteristics. Participants differed consistently in the strength of their attention to faces across images (Cronbach's \(\alpha = .88\)). A stronger preference for faces was correlated positively with openness to experience, extraversion, agreeableness and empathizing and was correlated negatively with social anxiety, depression levels and alexithymia. A dominant common factor, which may be best understood as general psychopathology, was correlated with attention towards faces at \(r = -.36\). Social anxiety stood out as only variable which explained attention towards faces above and beyond the Big Five personality factors. We compare these results with previous studies to assess the generalizability of social attention between laboratory and real-life situations and to discuss a possible bridging role of virtual reality.
9.15am
Self-focused attention in social anxiety: A naturalistic eye-tracking investigation
9.30am
Gaze and speech behavior in parent-child conversations
Gijs Holleman, Ignace Hooge, Jorg Huijding, Maja Deković, Chantal Kemner, & Roy Hessels
In this study, we investigated gaze behavior and its relation to speech during video-mediated face-to-face conversations between parents and their preadolescent children (8-10 years). 81 parent-child dyads engaged in two brief conversations about cooperative and conflictive family topics. We used eye tracking, audio-, and video-recordings to assess what regions of the face are looked at during episodes of speaking and listening, and whether patterns of gaze and speech were influenced by the topic of conversation. Regarding speech, our results show that children spoke more in the cooperation-scenario whereas parents spoke more in the conflict-scenario. Regarding gaze, we found that both parents and children looked more at the other's mouth region while listening compared to while speaking, which converges with previous studies about face scanning during speech perception. Furthermore, we found that parents gazed slightly more at the eyes of their children in the conflict-scenario compared to the cooperation-scenario. Results are discussed in terms of the role that parents and children take during cooperative and conflictive interactions, and how gaze behavior may support and coordinate such interactions.
Face and emotional perception
Face and emotional perception
Session chaired by Benedikt Wirth
10.30am
Affective and semantic incongruency of occupational labels have disparate effects both in supra- and subthreshold face recognition tasks
Ferenc Kocsor, Rita Kertész, Blanka Schaadt, & Tas Ferencz
Recognition accuracy of faces can be enhanced by stimulating a specific part of the face-space. This can happen with labels suggesting social status, occupation, or typical behavior. The effect might be ascribed either to semantic information playing a role in recollection of memories, affective information evoked by the labels, or both. We designed two studies in which participants learned associations between faces and occupational labels. In Study 1, subjects were exposed to faces, preceded by labels for 500ms. In Study 2, they were primed with labels for 33ms. Faces appeared after either the learned labels, or after incorrectly paired labels, intertwined with unfamiliar faces. In the incongruent conditions the semantic and affective distances between the presented and the correct labels varied. The task was to respond to those faces they have encountered in the learning phase. Subjects responded slower to incongruent stimuli in both studies. This confirms the assumption that false labels activate a specific region of the face-space, which puts a cognitive load on recognition. In Study 1, the joint difference of semantic and affective contents had the largest effect. In Study 2, in contrast, reaction time was mostly influenced by the affective distance, semantic difference did not have significant effect on face recognition speed. To conclude, semantic information has a strong effect when it is consciously evaluated because it is mapped onto a smaller region of face-space and provides more accurate information. However, the processing of emotions evoked by a label might be quicker and precede conscious responses.
10.45am
Impression formation based on the face is affected by the age of a person's voice
Helene Kreysa & Stefan R. Schweinberger
In natural contexts, person perception is multi-modal, with information from the face and voice being rapidly integrated to form a unified representation of the person in question (Young et al., 2020). Although many socially important aspects of a person can be inferred reliably from either modality, one modality may be used preferentially when both are available (Mileva et al., 2018). In order to tease apart differential contributions of faces and voices to first impression formation based on an individual's age, we combined photos of middle-aged faces (ca. 40 - 50 years of age; Ebner et al., 2010) with voices (Zäske et al., 2019) that were either younger or older than the faces (ca. 20 vs. > 65 years old). Participants were asked to rate each person for trustworthiness, attractiveness, and dominance, and, in a final block, to guess the person's age. Two experiments differed only with regard to the age of the participants themselves: Experiment 1 tested 27 students (\(M = 22\) years old); Experiment 2 tested 29 senior citizens (\(M = 75\) years). The students rated faces combined with younger voices as more attractive than the same faces combined with older voices; for senior raters, voice age did not affect attractiveness ratings. For dominance ratings, the pattern reversed: Students experienced faces combined with older voices as more dominant, senior raters found faces with younger voices more dominant. Trustworthiness ratings were not affected by voice age or participant age. Age guesses tended to be close to the actual age of the face. Interestingly, despite age differences of roughly \(\pm 20\) years between faces and voices, only very few face-voice pairings were experienced as strange or mismatching, which is an important finding with regard to multimodal integration.
11.00am
Response priming reveals a bias in trustworthiness evaluations of female versus male faces
Benedikt Wirth & Dirk Wentura
Using an innovative variant of the response-priming paradigm, ten Brinke et al. (2014, Psychological Science) seemingly found evidence for unconscious lie detection: Participants were asked to classify target words as truth-related or lie-related. These words were preceded by photographs of individuals whom the participants had previously watched being interrogated about committing a mock crime. Interestingly, pictures of those suspects who lied to the interrogator acted as lie primes and pictures of those suspects who told the truth acted as truth primes. However, a reanalysis of ten Brinke et al.'s data showed that this priming effect was most likely caused by a gender bias in facial trustworthiness evaluations. In order to further test the hypothesis that female faces are generally perceived to be more trustworthy than male faces, we conducted two response-priming experiments \(N_1 = 39\), \(N_2 = 61\). Participants were asked to classify target words as truth-related or lie-related. Each target word was preceded by a male or female face prime. An independent sample had explicitly rated half of these face primes as trustworthy, the other half as untrustworthy. We found that generally trustworthy faces acted as truth primes and untrustworthy faces acted as lie primes. Furthermore, within the class of trustworthy faces, female faces acted as truth primes and male faces acted as lie primes. However, this gender effect did not occur for untrustworthy faces. These results suggest that (a) spontaneous facial trustworthiness evaluations are biased by face gender and that (b) facial trustworthiness evaluations can be measured with a variant of response priming originally developed to measure unconscious lie detection.
11.15am
Involuntary evaluation of others' emotional expressions depends on expresser's ethnicity. Further evidence for the social message account by using the extrinsic affective Simon task
Emre Gurbuz, Andrea Paulus, & Dirk Wentura
If emotional expressions are presented in tasks that allow for involuntary assessment of valence, typically joy and fear expressions elicit positive and negative evaluations, respectively. If faces of different ethnicities are presented, typically evaluations conform to prejudices (ingroup = positive; outgroup = negative). What happens if faces are characterized by two features (i.e., emotion and ethnicity)? The social message account (SMA) assumes that the features will be integrated: ingroup joy indicates affiliation whereas outgroup joy indicates dominance; ingroup fear is a warning whereas outgroup fear signals submission. Evidence was found in an approach-avoidance task (AAT; Paulus & Wentura, 2014): Ingroup joy and outgroup fear facilitated approach whereas ingroup fear and outgroup joy facilitated avoidance. However, in the evaluative priming task (EPT) ‐ aiming at involuntary evaluation ‐ both features independently triggered priming effects (Paulus & Wentura, 2018). A difference between the tasks was that ATT required direct reactions to faces while in EPT faces served as task-irrelevant primes. Therefore, we examined the evaluation with the Extrinsic Affective Simon Paradigm (EAST; \(N = 120\), online). As in AAT, the EAST let participants react to faces directly. We used negative and positive words to load two keys with negative/positive valence, respectively. In other trials, happy and fearful faces of Middle Eastern and White Caucasian males were presented. Using the valent keys, participants had to categorize which side of the face was blurred. The affective meaning was indeed involuntarily processed and it conformed to SMA. That is, ingroup joy and outgroup fear facilitated negative responses whereas ingroup fear and outgroup joy facilitated positive responses.
11.30am
Conflict or intentions: Affective mismatch not social intentions explains reactions to emotions of ingroup and outgroup members
Robert Huber, Julia Kozlik, & Rico Fischer
A frequent observation in intergroup research is that reactions to emotional displays differ as a function of group membership. A confrontation with a happy ingroup member, for example, results in concordant approach reactions, whereas happy outgroup members induce discordant avoidance reactions. There are two explanatory approaches for this phenomenon: (1) The social intentions account proposes that this response divergence is determined by underlying social messages within the face. (2) The processing conflict account assumes this response pattern to be caused by an affective conflict triggered by incongruent combinations of emotion and the affective connotation of group membership. Our goal was to distinguish between both approaches by separating social intentions from the affective conflict. For this, we created a new paradigm, in which the participant is an outside observer of a social interaction between two displayed persons (i.e., a target and a distractor person). Target and distractor could either share their group membership or not. Participants were required to categorize the emotion displayed by the target, which also could either be ingroup or outgroup in relation to the participant. Thus, with this setup, social intentions critically hinge on the target-distractor group relation, whereas affective conflicts result from the target-participant relation. In multiple experiments, the target-participant relation interacted with the emotion expressed by the target, suggesting that responses are affected by affective conflict. However, the target-distractor relation did not interact with emotion, which indicates no influence of social intentions on participants' reactions. In sum, our data speak in favor of the conflict instead of the social intentions account.
Emotions and affect
Emotions and affect
Session chaired by Timea Folyi
4.00pm
Fast and involuntary interactions between emotional vocalizations, emotional music, and facial emotions
Timea Folyi & Dirk Wentura
In a series of experiments, we investigated the influence of brief auditory affective context on visual affective processing to test possible underlying processes and their affective specificity. We used a multimodal variant of the evaluative and emotion priming paradigm (for the latter, see e.g., Rohr et al., 2012), and contrary to typical multimodal studies, we used brief sound durations (200‐600ms) and stimulus-onset-asynchronies (0‐200ms) in order to preserve the fast and involuntary character of the paradigm. Our first results showed emotion-specific priming effects between non-verbal vocalizations and facial expressions. As there is plausibly a strong associative relation between emotional vocalizations and facial expressions due to natural audiovisual co-occurrence and semantic relatedness, we tested whether these effects generalize to emotional music and faces. We presented brief (600ms long) excerpts of joyful, sad, anger or fear-related instrumental music as primes and emotional facial expressions as targets. Contrary to what is suggested by previous findings (e.g., Marin et al., 2012), there was an emotion-specific influence of brief musical context on the categorization of visual emotions. These results suggest that emotions that are automatically (i.e., fast and involuntary) conveyed by music can be represented similarly to social-communicative emotion categories in contrast to only arousal and/or valence. Furthermore, the study indicates that emotion-specific cross-modal interactions are not limited to strongly associated sounds and visual stimuli. Two further experiments varying the task provided hints that the underling processes are not limited to response bias; the possible processes will be discussed based on these experiments.
4.15pm
Semantic or affective primacy? Perceptual latencies of stimulus recognition and affect for sounds, measured with the rotating spot method
Philipp Franikowski & Rainer Reisenzein
According to the semantic primacy hypothesis of emotion generation, stimuli must be recognized (semantically categorized) to evoke emotions. Chronometric evidence for this hypothesis has been obtained in previous studies for pictures as affect-evoking stimuli. Here, we tested whether these findings generalize to affect-evoking sounds. To measure the perceptual latencies of sound recognition and sound-evoked affect, we used the rotating spot method of timing subjective events. Participants listened to 12 pleasant and 12 unpleasant sounds familiar from everyday life while watching a spot that revolved quickly around the perimeter of a simulated clock face. In different blocks of trials, they indicated when they experienced the pleasant or unpleasant feeling evoked by a sound, and recognized the sound, by reporting the position of the spot at the time when these mental events occurred. Replicating previous findings obtained for affective pictures, the latency of object recognition was shorter than the latency of affect, and the two latencies were positively correlated. The findings extend chronometric support for the semantic primacy hypothesis to affective sounds.
4.30pm
Semantic or affective primacy? Latencies of object recognition and affect: A comparison of rotating spot and speeded reaction time measures
Rainer Reisenzein & Philipp Franikowski
According to the semantic primacy hypothesis of emotion generation, stimuli must be recognized (semantically categorized) to evoke emotions. Chronometric evidence for this hypothesis has been obtained in previous studies by the authors with several different latency measurement methods: temporal order judgments, simultaneity judgments, the rotating clock hand (RH) method, and speeded reaction time (sRT) measurements. To consolidate, generalize and extend these findings, we conducted a new study in which RH latency measurements were directly compared to sRT measurements obtained from the same participants. An improved RH method was used and two clock hand speeds were compared. In sessions one and two, the participants viewed 40 pleasant and 40 unpleasant pictures while a clock hand, symbolized by a colored line, revolved quickly around the center of the picture. In different blocks of trials, the participants indicated when they experienced the pleasant or unpleasant feeling evoked by the picture, and recognized the depicted objects or scene, by reporting the position of the clock hand at the time when these mental events occurred. In a third session, the same latency measurements were obtained using sRT. In addition, we timed the recognition of color and the speed of valence judgments using the two latency measurement methods. Currently we are still in the process of data collection.
4.45pm
Intraindividual variability in affect: A formalized, theoretical approach
Maria Wirth, Andreas Voss, & Klaus Rothermund
Studying the ebb and flow of affect in daily life provides important insights into psychological functioning and well-being. However, little attention has been paid to the sources and underlying processes of these short-term changes. We propose a model in which affect is conceptualized as the output of dynamic processes. Given that affect reflects transactions between an organism and its proximal environment, we relate the evaluative aspect of events in terms of their pleasantness or unpleasantness (iV, "valence") to fluctuations in momentary affective experience (dV, "affect"). The model explains affective experience as resulting from the cumulative effects of previous valent events ("accumulation principle"). The core of the model consists of parameters that moderate the relation between valent events and affective experience. These parameters reflect individual differences in the extremity of short-term changes (reactivity) and in longer-term changes in affective experience (attenuation) caused by positive or negative events. A simulation study revealed identifiability of the model's core parameters via Bayesian modeling. An empirical application of our model will be presented using daily affect and event ratings from 315 individuals ranging in age between \(14-86\) years. The general pattern of results suggested age-related improvements. Older adults were less reactive to hassles and regulated hassle impact more strongly than younger adults. Affect elicited by uplifts was more stable for older than younger adults. We discuss potentials and limitations of the approach and close with an outlook on the broader implications for understanding emotional development.
5.00pm
Are you hurt yet? An interactive approach to reactive aggression
Vanessa Mitschke & Andreas Eder
Consequences of enacting revenge range from the often-devastating escalation of violence to experiencing a calm sense of forgiveness. Research on revenge often treats vengeful acts as singular experiences, an approach which fails to account for the role of outcome satisfaction in the regulation of reactive aggression. In a series of studies, we showcase the potential of integrating target reactions into dynamic revenge sequences to investigate the paradoxical consequences of revenge. In treating the provoked avenger as seeking a desired state within their interaction partner (such as defeat, pain, compliance), we discuss the complex affective evaluations of revenge outcomes, shedding light on mixed findings concerning the presence of negative and positive affect following revenge punishments. In Studies 1‐2, we investigated affective responses towards negative target reactions and the regulation of aggression based on the evaluation of distinct harm indicators. In Study 3, we observed the significant escalation potential of opponent states incongruent with the desired opponent state, as it is the case when being confronted with Schadenfreude. In an additional series of studies, we observed changes in memory of opponent faces following vengeful encounters. Based on target state indicators (pain, anger), face memory was distorted, resulting in less negative representations of opponents that expressed pain. We argue that integrating dynamic interaction aspects, such as facial expressions as indicators of the target state, aids the process of identifying aggression trajectories and furthers our understanding of the goals and motives present during the escalation or de-escalation of vengeful acts.
Mauve Atrium
Metamemory
SymposiumMetamemory
Symposium organized by Beatrice G. Kuhlmann and Monika Undorf
Metamemory research focuses on people's knowledge about learning and memory and on the assessment (monitoring) and regulation (control) of their ongoing learning and memory processes. This symposium presents current research on metamemory from labs in Poland, Portugal, Turkey, and Germany. The first two talks demonstrate the specific sensitivity of metamemory judgments in established memory paradigms. Using a misinformation paradigm, Yüksel and Besken show that degraded viewing conditions reduce both memory predictions and actual memory performance, but the integration of misinformation to retrieval remains unaffected. Hanczakowski et al. demonstrate that metamemory judgments disclose modulating effects of the number of context-to-item associations on context reinstatement even when the respective effects on recognition memory are inconclusive. The next two talks focus on the basis of people's predictions of their future memory performance (judgments of learning, JOLs). Mendes and Undorf present research indicating that JOLs rely on objective and subjective word frequency in multiple-cue situations. Undorf, Navarro-Báez, and Bröder present a novel methodological approach for uncovering hitherto elusive influences of idiosyncratic information on JOLs. The final two talks investigate the debiasing of metamemory illusions. Schaper et al. report evidence that alleviating the expectancy illusion improves the accuracy of metamemory monitoring and control but has little effect on actual memory performance. Zimdahl and Undorf report results showing that neither warnings nor incentives reduce the hindsight bias on JOLs. Together, the six talks of this symposium present new insights on metamemory monitoring and control.
8.30am
The effects of perceptual disfluency on predicted, actual and false memory for a misinformation paradigm
Ezgi Melisa Yüksel & Miri Beskenpresenting
Perceptual disfluency can affect both metacognitive judgments and actual memory performance. In a set of two experiments, we investigated how participants may integrate information from a misinformation phase when they encountered perceptually disfluent information in the original encoding phase for witnessed events. In the original encoding session, participants were exposed to eight stories with fifty images each. Half of the stories were presented in a fluent condition with intact pictures, whereas the other half were presented in a disfluent condition, masked with a checkerboard pattern. In the misinformation phase, they were shown all the stories again in their intact form with some of the details changed. In the test session, participants completed a multiple-choice recognition memory test, retrieving details from the original encoding phase. For Experiment 1, the results revealed more accurate retrieval for fluent than disfluent stories. Moreover, the misinformation manipulation created a significant amount of false memories; yet, the effect of the perceptual disfluency manipulation on integration of misinformation to the original story was not significant. For Experiment 2, participants also provided metacognitive memory judgments in both encoding and misinformation phases. In terms of actual and false memory performance, the findings were similar to Experiment 1. In terms of metacognitive judgments, participants produced higher memory predictions for fluent than disfluent stories, in line with their actual memory performance. The current study might have implications for eyewitness memory for both actual and predicted memory performance, especially when participants cannot perceive the information at original encoding phase clearly.
8.45am
Context fan in recognition: Insights from metamemory measures
Maciej Hanczakowski, Oliwia Zaborowska, & Katarzyna Zawadzka
Context reinstatement effects in recognition are often small and difficult to replicate across experiments. One solution to this problem has been to use metacognitive judgments as measures that can reveal the impact of context on recognition in the absence of an effect on recognition performance. Here we present three experiments that used this metacognitive approach to examine an old issue regarding context reinstatement in recognition: the extent to which this effect is modulated by the number of context-to-item associations. Experiment 1 assessed the role of context fan for old/new recognition, while Experiment 2 employed forced-choice recognition, with both tests accompanied by retrospective confidence judgments. Recognition performance measures proved equivocal across experiments with regard to both context reinstatement and the modulating role of context fan, while the metacognitive measure revealed clear effects of both. Experiment 3 extended these results to spontaneous recognition during restudy as assessed by judgments of learning. Together, the results reveal robust effects of both context reinstatement and the modulating effect of fan, confirming the usefulness of metacognitive measures for revealing memory effects.
9.00am
Objective and subjective word frequency affect judgments of learning in multiple-cue situations
Pedro Simão Mendes & Monika Undorf
Research has shown that people are capable of using multiple cues when making metacognitive judgments such as judgments of learning (JOLs) ‐ predictions of one's future memory performance. Past research has shown that word frequency (i.e., how often words are encountered in the language) affects people's JOLs when manipulated in isolation. It remains unclear, however, whether subjective word frequency (i.e., the subjective familiarity of words) would impact JOLs over and above objective word frequency and whether frequency effects on JOLs would persist when another cue is simultaneously manipulated. Here, we aimed to close these gaps by testing if objective and subjective word frequency affect JOLs when font size is also manipulated. In three experiments, participants studied words that varied in word frequency (Experiment 1: high and low objective frequency; Experiment 2: a continuum ranging from high to low objective frequency; Experiment 3: high and low subjective and objective frequency) and were presented in a large (48 point) or a small (18 point) font size, made immediate JOLs for all words, and completed a free recall test. Results showed that people based their JOLs on word frequency and font size even though only word frequency impacted their actual memory performance. Moreover, both objective and subjective frequency guided people's JOLs, suggesting that subjective word frequency contributes to the word frequency effect in metamemory. We conclude that word frequency effects in metamemory are pervasive in multiple-cue situations.
9.15am
Uncovering idiosyncratic influences on judgments of learning
Monika Undorf, Sofia Navarro Báez, & Arndt Bröder
Numerous studies showed that people base predictions of their own future memory performance—judgments of learning—on general cues pertaining to the study materials (e.g., word frequency) and learning conditions (e.g., number of study opportunities). In contrast, the impact of idiosyncratic information such as the personal significance of items on judgments of learning has remained untested. We propose that idiosyncratic influences on metamemory can be measured by the nonlinear component C of Egon Brunswik's (1952) lens model. In two experiments, we made randomly chosen items personally significant (Experiment 1) or assessed the personal significance of items (Experiment 2). Personal significance increased judgments of learning and memory performance. Including personal significance as a predictor in the lens model reduced C, whereas including previous encounters with items in the experiment did not. Hence, the lens model's C parameter captures idiosyncratic influences on metamemory judgments and may serve as a useful tool for future research.
9.30am
Remedying the metamemory expectancy illusion in source monitoring: Are there effects on restudy choices and source memory?
Marie Luisa Schaper, Carolin Hey, Ute J. Bayen, Timo Asbeck, Hannah Rodloff, & Mara von Schwartzenberg
In source monitoring, schema-based expectations bias metamemory monitoring and control (Schaper & Bayen, 2021). Source memory is better for unexpected than expected sources (e.g., "hairdryer in the kitchen" vs. "oven in the kitchen", inconsistency effect). However, people predict the opposite: They show an expectancy illusion. This illusion biases restudy choices. Participants prefer to restudy unexpected source‐item pairs, even though source memory is already quite good for these pairs. In the current study, we aimed to remedy the expectancy illusion (Schaper et al., in press) and debias restudy choices by implementing a delay between study and metamemory judgments. We further asked whether debiased restudy choices enhanced source memory for expected pairs, thereby reducing the inconsistency effect. Two participant groups (\(n = 72\) each) studied expected and unexpected source‐item pairs. They provided metamemory judgments and selected pairs for restudy either at study (immediate) or after a delay (delayed). They then restudied the selected pairs and completed a source-monitoring test. The immediate group predicted better source memory for expected pairs and selected more unexpected pairs for restudy. The delayed group predicted a null effect on source memory, and selected equal numbers of expected and unexpected pairs. Thus, delaying judgments partially remedied the expectancy illusion in metamemory monitoring and debiased metamemory control. Despite differential restudy choices, both groups showed equally strong inconsistency effects on source memory. Debiasing metamemory monitoring and control via delayed judgments only weakly affected source memory.
9.45am
The hindsight bias on judgments of learning is not reduced by warnings or incentives
Malte F. Zimdahl & Monika Undorf
Once people acquire outcome knowledge, they tend to overestimate in hindsight what they knew in foresight. This hindsight bias could be demonstrated for a variety of judgments and could recently also be shown for judgments of learning (JOLs). The two current experiments tested whether the hindsight bias on JOLs can be reduced by two different debiasing methods: warnings and incentives. Both experiments started with a learning phase, in which participants studied word pairs and made JOLs. Afterwards, participants acquired outcome knowledge for half the items by completing a memory test prior to recollecting their original JOLs. In Experiment 1, one group of participants received a detailed description about the hindsight bias on JOLs and instructions to avoid it before recollecting their JOLs, whereas another group did not. In Experiment 2, we incentivized correct JOL recollections by monetary bonuses. Both experiments revealed that, for word pairs with outcome knowledge, recollected JOLs were closer to people's actual memory performance. This hindsight bias on JOLs was unaffected by warnings and incentives, with Bayesian analyses providing evidence that neither debiasing method attenuated the hindsight bias on JOLs. These findings suggests that the hindsight bias on JOLs is due to cognitive rather than motivational processes.
Metamemory and long-term memory
Metamemory and long-term memory
Session chaired by Arndt Bröder
10.30am
Metamemory accuracy for scene pictures: It depends on how and when you ask
Arndt Bröder, Sofia Navarro Báez, & Monika Undorf
Although it is well known that recognition memory for naturalistic photographs of scenes is amazingly good and consistent across different observers (i.e., scenes are consistently likely to be remembered or forgotten), less is known about the accuracy of metamemory judgments predicting the stimulus-specific memory performance. There are conflicting results in the literature reporting either a lack of predictive validity of judgments of scene memorability (Isola et al., 2014) or a reasonably good metamemory accuracy of judgments of learning (JOLs; Undorf & Bröder, 2021). The former was assessed by asking for the generic stimulus attribute "memorability" (memorability judgment, MJ), whereas in a JOL procedure, participants predict their own later memory performance for a stimulus. In two experiments (overall \(N = 102\)), we directly contrasted the two ways of assessing metamemory in a within-subjects-design to test whether the response format was responsible for the difference in accuracy. In both experiments, we found an interaction of response format with task order, showing much improved MJs after a preceding JOL task whereas MJs were inferior to JOLs when administered first. This robust result suggests that the assessment of the general stimulus characteristic "memorability" can be improved by a preceding learning experience during which participant predict their own memory performance for different stimuli.
10.45am
Judgements of learning are based on memory strength: Evidence from visual and verbal tasks
Julia Krasnoff & Alessandra S. Souza
The theory that Judgements of Learning (JOLs) are based on memory strength is considered outdated. However, this literature ignored the possibility that the strength of working memory representations is used to infer the likelihood of remembering at a delayed memory test. In three preregistered experiments, we show that working memory representations guide inaccurate JOLs. In Experiment 1, participants memorized sequences of two or four colored objects, then they provided JOLs for a delayed test, and performed an immediate test on the objects' colors. After learning 200 objects, the delayed test on all studied objects followed. Sequence-length affected immediate, but not delayed performance. JOLs, however, were higher for sequences of two than four, correlating higher with immediate than delayed performance. Experiment 2 replicated the sequence-length effect on JOLs in the absence of an immediate memory test. In Experiment 3, we observed that working memory strength guided JOLs in a verbal task: Participants memorized sequences of six words, provided JOLs for each word, and subsequently performed an immediate memory test on the words. A final delayed memory test occurred after learning of 48 words. Immediate memory performance was highest for words presented in the first positions of the list and decreased over serial position. In line with the memory strength theory, JOLs were also higher for the first list positions and decreased with serial position. Our results suggest that people base their JOLs on the quality of their ongoing working memory representations, therefore challenging the predominant view of exclusively inferential, cue-based JOLs.
11.00am
Monitoring accuracy and working memory: A competition for resources account
Donna Bryce & Florian Kattner
Knowing what one knows and accurately monitoring one's own capacities and performance on a moment-to-moment basis are important determinants of both immediate task success and educational achievement. Individual differences in such metacognitive monitoring are well documented, but what determines an individual's monitoring accuracy in a particular context is yet to be fully understood. One candidate contributor to monitoring accuracy is working memory. In this study, we investigated whether and how working memory contributes to the accuracy of monitoring processes. Up until now, most evidence for a positive relationship between working memory and monitoring accuracy has been provided by correlational studies. Here, an experimental approach was applied in which the working memory demands of the task were varied and the impact on monitoring accuracy assessed. In two out of three experiments, monitoring accuracy suffered when working memory demands increased. These results indicate that the overall demands of the task determine whether participants engage in more effortful direct online monitoring or rather integrate various cues to generate confidence judgements. Further, it appears that the processes involved in direct online monitoring compete for resources with working memory.
11.15am
Value-directed forgetting: Are negatively valued words (and their values) inhibited?
Ryan Patrick Hackländer, Ann-Kathrin Rolke, & Helge Schlüter
The notion that humans can voluntarily forget some information of their choosing has a long theoretical history in the field of psychology. Since the 60's there has also been a strong empirical history of testing voluntary forgetting in the laboratory. The method of choice has been to use directed forgetting experiments, in which researchers tell subjects which items to remember and which to forget. While the finding that subjects have a harder time later remembering information they were initially told to forget is quite robust, it is questionable whether this "forgetting" is truly voluntary. In the current experiment, we utilize a variation of the value-directed remembering paradigm and assign values to different, to-be-remembered words. The values are either positive (+5 points or +10 points), negative (-5 points or -10 points), or neutral (0 points). In a study phase, subjects first encode a word and then are presented with the cue as to how many points they will receive if they later report the word on a free recall test. However, during the actual recall test, the subjects are told that all words, regardless of prior cue, are worth +1 point, and they should try to maximize their point value. With this method we are interested in a) replicating previous findings showing worse memory for negatively valued words, b) determining if source memory (for the associated point value) is also reduced for negatively valued words (which would be a further measurement of forgetting), and crucially c) determining whether negatively valued words are remembered at lower rates than neutral valued words, which would potentially show that negatively valued words are not only not selectively rehearsed, but possibly actively inhibited.
11.30am
Semantic associative effects in directed forgetting: Evidence from ERPs
Yutong Li, Xue Sui, Markus Hofmann, & Ralph Roger Radach
Can humans intentionally "forget" implicitly conditioned associations between words that have been shaped due to a long common learning history? In the present study, we use co-occurrence statistics to define associations, and examine target words with many vs. few associated words in a directed forgetting task. 28 participants learned a series of Chinese words with many or few associated words in the study phase. Each target word was followed either by cue R prompting the subjects to remember the word, or cue F prompting the participants to forget the word. In the test phase, participants made an old/new word decision on presented words. The behavioral results showed that the accuracy of target words with many associates was higher, which suggests that these implicitly conditioned long-term associates can help to follow the R/F instruction. Words with only a few associates evoked a larger P2 and LPC amplitude in the study phase, potentially indicating a deeper processing of associatively isolated words. In an N2 time window, F cues evoked a larger amplitude than R cues, reflecting the recruitment of cognitive (memory) control. As opposed to the study phase, many associates elicited a larger P200 amplitude, probably because of their shallower encoding. In the frontal region, in contrast, words with only a few associates elicited a larger FN400 amplitude only for erroneously remembered items, suggesting a familiarity-based false memory decision. In sum, our results suggest that intentional forgetting and remembering is affected by a common learning history of the study items, highlighting the importance of long-term associations for the cognitive control of episodic memory processes.
11.45am
Wakeful resting effects on long-term memory are moderated by individual differences in working memory capacity
Robert Marhenke, Pierre Sachse, & Markus Martini
Wakeful resting after encoding new information can support memory retention, while task-related cognition interferes with long-term memory consolidation. Recent evidence suggests that individual differences in working memory capacity (WMC), a factor believed to contribute substantially to individual differences in "online" processing, encoding and retrieval of information, contribute to differences in "offline" memory consolidation during sleep. We tested whether individual differences in WMC also moderate the effect of a brief period of wakeful resting compared to performing a distractor task subsequent to encoding a word list. Ninety-eight participants encoded and immediately recalled a word list followed by either an 8 min wakeful resting period (eyes closed, relaxed) or by performing an adapted version of the d2 test of attention for 8 min. At the end of the experimental session (after 12‐24 min) and again, after 7 days, participants were required to complete a surprise free recall test of both word lists. Our results suggest that individual differences in WMC are a central moderating factor for the effect of post-learning activity on memory retention. While we did not find a universal difference in memory retention between conditions, we found a cross-over interaction between word retention in each condition and participants' WMC over the retention interval of 12-24 min, as well as over 7 days. While retention increased for higher WMC individuals in the resting condition, higher WMC was negatively related to word retention in the distractor condition. Our results shed new light on contradictory findings regarding the effect of post-learning wakeful rest on memory retention and show the importance of considering individual differences in memory experiments.
Verbal long-term memory
Verbal long-term memory
Session chaired by Jan Philipp Röer
In this talk, I will give a brief overview of three experiments that investigated the effects of background speech on subsequent behaviour. The procedure was the same in all experiments. Participants performed a serial recall task, either in silence or while task-irrelevant background speech was presented over headphones. Subsequently, participants undertook what they believed to be an unrelated norming study, in which they were asked to write down exemplars from different semantic categories (Experiment 1), solve CRAT problems (Experiment 2), and rate how confident they were that a statement was true or false (Experiment 3). In all three experiments, a clear effect of background speech on subsequent behaviour was observed. Previously presented words were more likely to be written down as category exemplars (Experiment 1), they were more likely to be used as solution words for CRAT problems (Experiment 2), and subjective truth ratings were higher for previously presented statements compared with new statements (Experiment 3). Theoretical implications and avenues for future research are discussed.
4.15pm
Effects of irrelevant speech and word length on immediate serial recall: What role of rehearsal?
Abdullah Jelelati, Larissa Leist, Thomas Lachmann, & Maria Klatte
Visual-verbal serial recall is reliably impaired by task-irrelevant background speech that the participants are instructed to ignore ("irrelevant speech effect", ISE). The wordlength effect (WLE) is the serial recall advantage for short when compared to long words. The ISE has been attributed to specific interference of obligatory sound processing with either the integrity of phonological traces in a phonological short-term store (phonological loop account), or the efficiency of the rehearsal process employed for serial order retention (changing-state account). Both the phonological loop and the changing-state account assume that the WLE reflects articulatory rehearsal of the list items, and that the ISE and the WLE exert independent, additive effects on serial recall performance. However, the role of rehearsal in the ISE and WLE has been questioned by studies showing that the effects survive when rehearsal is blocked through rapid presentation of the list items. In a series of current experiments, we investigate the combined effects of irrelevant speech and word length on serial order reconstruction of German nouns presented pictorially, using standard and rapid presentation rates. Serial order reconstruction was used instead of serial recall in order to rule out an impact of recall time, and thereby output interference, in the word length effect. With a standard presentation rate, we found reliable and independent effects of wordlength and irrelevant speech. These results will be compared to those from current experiments with a rapid presentation rate. Implications for dominant accounts of the ISE and WLE are discussed.
4.30pm
Differential effects on serial recall of verbal and spatial items in children and adults due to irrelevant speech
Larissa Leist, Markus Georgi, Sabine J. Schlittmeier, Thomas Lachmann, & Maria Klatte
Task-irrelevant speech, which the participants are instructed to ignore, interferes with immediate serial recall of visually presented items. The so-called Irrelevant Speech Effect (ISE) has been linked to attentional capture, phonological interference, and interference with serial order representations of both verbal and visuospatial items. We investigated the effects of background speech on serial order reconstruction of verbal and spatial items in third-grade children and adults in order to learn more about the roles of phonological processing, serial order retention, and attention control in the ISE. In the verbal task, the children (adults) were shown picture sequences of 5 (8) German nouns (e.g., frog, bed, moon). In the spatial task, sequences of 5 (7) dots were presented at different screen locations to the children (adults). In the recall phase of both tasks, all items were displayed on the screen at the same time, and participants had to mouse-click on the items in the right serial order. In both age groups, we found significant performance decrements due to irrelevant speech in the verbal task. Children were more impaired than adults. Irrelevant speech had no effect on the spatial task performance of either children or adults. These findings suggest that phonological processing is a precondition for ISE evocation.
4.45pm
When is semantic auditory distraction beneficial to memory performance ?
Krzysztof Piatkowski, Katarzyna Zawadzka, & Maciej Hanczakowski
With overheard speech comes great distractibility. Yet, some forms of verbal distraction are capable of improving rather than harming our cognitive performance. While distraction consisting of spoken words semantically related to to-be-remembered targets is usually detrimental to memory, this pattern is reversed when distractors can uniquely specify the category of a particular target. Current theoretical explanations of this phenomenon focus on the role of distractors in elaborating targets' representations in memory. It is not clear, however, whether such elaboration is automatic or requires the engagement of attentional resources. We investigated this problem by varying the engagement of attention during learning under distraction with a manipulation of perceptual difficulty of processing targets, assumed in distraction studies to foster greater task immersion. We presented participants with lists of visual targets along with auditory distractors and manipulated target-distractor semantic relatedness as well as target's processing difficulty ‐ by presenting to-be-remembered words either upright or inverted. In Experiment 1, we manipulated processing difficulty across study lists to vary attentional engagement across broad chunks of the study activity. In Experiment 2, the same manipulation was implemented on an item-by-item basis within a study list. The results of Experiment 1 replicated the phenomenon of benefits of semantically related distraction for memory and this effect emerged independently of the level of processing difficulty. These results suggest that the process of semantic elaboration based on information conveyed by auditory distraction is largely independent of the allocation of attentional resources.
5.00pm
Animacy improves recognition by enhancing recollection: Convergent evidence from two experimental procedures
Gesa Fee Komar, Laura Mieth, Axel Buchner, & Raoul Bell
Animacy has been established as an important determinant of memory during the last decade. Considering its postulated evolutionary-adaptive significance, the memory advantage of words for animate entities over words for inanimate entities (henceforth animacy effect) should not only affect the quantity but also the quality of remembering. We tested this assumption by examining animacy effects both on subjectively experienced retrieval states using the remember-know-guess paradigm (Experiment 1) and on objective performance-based measures of the quality of recognition using the process-dissociation procedure (Experiment 2). Proximate accounts of the animacy effect that focus on elaborative encoding and attention allow predicting animacy to selectively enhance detailed recollection but not the acontextual feeling of familiarity. We adopted a multinomial modeling approach to disentangle recollection, familiarity, and different types of guessing processes. Results obtained from the two experimental procedures convergently show that animacy selectively enhances recollection but affects neither familiarity nor guessing. The recognition advantage of words for animate entities over words for inanimate entities is thus primarily driven by enhanced recollection. The results support the richness-of-encoding account and the attentional account of the animacy effect on memory.
Chestnut Chamber
Interplay between prior knowledge with new learning and memory
SymposiumInterplay between prior knowledge with new learning and memory
Symposium organized by Iryna Schommartz and Yee Lee Shing
Prior knowledge provides mental schemas and frameworks that promote efficient incorporation of new information during learning and benefit subsequent memory. At the same time, when new information does not fit into pre-existing schemas, it has been shown to both hinder and benefit learning and memory. In this symposium with six talks, we provide empirical insights about the interplay between learning/memory and prior knowledge, using experimental paradigms, virtual reality, functional magnetic resonance imaging, eye tracking techniques, as well as neuro-pharmacological approach. We showed that (i) when remembering recent experiences, semantic substitution may lead to memory bias depending on the task-relevance; (ii) the (in)congruency effect differentially impacts memory accuracy and precision and follows different temporal dynamics in the developing brain; (iii) prediction error as a result of mismatch between sensory evidence and prior knowledge may boost memory; (iv) schema-congruency determines the susceptibility of early memory formation and consolidation to the modulation by general anesthesia; (v) new schemas formed either through active agency or from an observer's perspective are associated with different neural correlates and memory malleability; (vi) pre-existing schemas of episodic information accumulation during active virtual navigation enhance long-delay memory consolidation and follow distinct developmental trajectories. Together, the presentations of this symposium provide empirical evidence that the effects of (in)congruency of new information on new learning and memory is complex and depends on several important modulating factors (i.e. developmental status, agency, schema strength, situation relevance, prediction error).
8.30am
Where is the toaster? Interplay of episodic memory and semantic knowledge during remembering of past events
Nicole Klein, Carina Zöllner, Sen Cheng, Ricarda I. Schubotz, Nikolai Axmacher, & Oliver T. Wolf
Misremembering details of recent experiences is ubiquitous in everyday life. Over the last years investigation prior knowledge as a contributing factor to memory biases became more prominent in memory research. Yet the exact underlying mechanisms remain unclear. The scenario construction model postulates that during encoding, only the gist of an episode is stored in the episodic memory trace and during retrieval, any needed information that is missing from that trace is substituted from semantic knowledge. In our study we aimed to investigate semantic substitution in a realistic, yet controlled setting. Participants navigated through a flat in a virtual reality (VR) and interacted with task relevant common household objects. The household items appeared in expected and unexpected locations to create conflicts between episodic memory and semantic expectations. During subsequent memory recall, we asked for the remembered location of an object and thus, we were able to distinguish between correct episodic recall, semantic substitution and guessing. In addition, we controlled the effect of attention, by manipulating task-relevance of objects. Both congruency and task-relevance predicted correct episodic memory retrieval. During recall, semantic substitution was more likely than guessing and occurred more frequently for task irrelevant objects. Our findings support the predictions of the scenario construction model. The new VR-paradigm is a promising tool for investigating semantic substitutions.
8.45am
Congruency effects memory on consolidation follow different temporal dynamics across development
Iryna Schommartz & Yee Lee Shing
Events that are congruent or incongruent with prior knowledge may benefit memory in different ways. However, little is known about (i) how the memory consolidation of events that are congruent or incongruent with prior knowledge evolves over time and (ii) how the temporal dynamics of this process may differ in children in comparison to adults, due to the ongoing development of the memory systems and accumulation of knowledge. In this study, we examined how short-delay (one night after encoding) and long-delay (two weeks after encoding) memory consolidation of events that are either congruent or incongruent with pre-existing knowledge evolves in 6-7-year-old younger children, 9-10-year-old older children and young adults. Memory was quantified with memory (i) accuracy and (ii) precision for object location. Based on the results so far (70% data collection completed), we found no age group difference in short-delay memory consolidation based on memory accuracy, regardless of congruency. However, younger children showed better short-delay memory consolidation for positioning precision of incongruent items in comparison to other age groups. Long-delay memory consolidation for incongruent items was lower than for congruent items based both on accuracy and precision across all age groups. Specifically for memory accuracy, younger children showed more efficient long-delay memory consolidation for congruent items but worse long-delay memory consolidation for incongruent items in comparison to adults. No such difference was observed between older children and adults. Taken together, the results suggest that memory consolidation of events (in)congruent with prior knowledge follows different temporal dynamics in different age groups regarding memory accuracy and precision.
9.00am
Understanding the memory-prediction error-liaison. An eye tracking study
Isabelle Ehrlich, Javier Ortiz-Tudela, & Yee Lee Shing
The reliable anticipation of the environment is one of the main tasks our brain performs continuously for us to deal with this complex world. In this respect, events that are not congruent with prior knowledge and derived expectations seem to be very important. Growing research on predictive processing suggests that a mismatch between knowledge-based predictions and sensory evidence, i.e., prediction error, can have a beneficial effect on the encoding and retrieval of the mismatching information. However, this positive effect on memory seems to be highly specific and the underlying mechanisms are not fully understood yet. The current eye-tracking study aims at further illuminating the precise requirements of such memory boosting mismatches by focusing on the multiple facets of declarative memory and the unsolved role of pattern separation as a possible mediator. In an adaptation of the classic mnemonic similarity task, we ask participants to form predictions during encoding that are eventually confirmed or violated. In a subsequent surprise memory task, participants are presented with (in)congruent targets, (in)congruent lures, neutral objects, and foils, and indicate whether the objects are old, similar, or new. Additionally, we test for memory of the specific encoding context as well as the timepoint in which the object was encountered during the experiment. Relating pupil response patterns to these various response combinations as well as lure discrimination indices, helps us to draw a clearer picture of the complex relationship between prediction error and declarative memory.
9.15am
Modulation of memory for schema-congruent and -incongruent information by propofol
Lukas Risse, Yee Lee Shing, Lili Landerer, Deetje Iggena, Patrizia Maier, Laura Hanke-Vela, Daniel Salchow, Mario Menk, Carsten Finke, & Christoph J. Ploner
Evidence mainly from animal studies shows that the relatedness of new information to prior knowledge promotes accelerated integration of new information into neocortical networks, modulating memory consolidation. Whether this is also associated with a shorter involvement of the hippocampus during consolidation remains to be elucidated. Here, we combined a variant of a classic schema-learning task with a neuro-pharmacological approach to transiently interfere with hippocampus-dependent steps of memory formation during the post-encoding period. Patients without neuropsychiatric comorbidities that undergo short general anesthesia for elective surgery of strabism were studied. Immediately prior to anesthesia, subjects learned, across three cycles, a word list that consisted of 40 words that were either congruent with a soundscape provided by headphones (restaurant noise) or completely unrelated. Subjects then received anesthesia with intravenous propofol (bolus of 3 mg/kg for induction, followed by 6 mg/kg/h for about 60 minutes for maintenance). Three hours later subjects were tested for recognition and recall of words. Although there was no effect of propofol on recognition memory of words, we found a clear general benefit for schema-congruent words in recall, consistent with facilitated consolidation of schema-congruent items. Propofol patients performed inferior to control subjects in recall of both word types, but also showed impaired learning, suggesting a combined effect of pre-surgical arousal and general anesthesia that preferentially affects schema-incongruent words. We infer that schema-congruency determines the susceptibility of early memory formation and consolidation to the modulation by general anesthesia.
9.30am
When I (do not) see myself: The role of agency and perspective in episodic memory
Benjamin Jainta, Sophie Siestrup, Nadiya El-Sourani, Ima Templer, Moritz Franz Nikolaus Wurm, Markus Werning, Sen Cheng, & Ricarda I. Schubotz
Common sense suggests that learning-by-doing is more effective than learning by observing to remember the relevant action episode later. Also, episodes can be recalled from the first-person (1pp) or the third-person perspective (3pp). Despite previous research showing that self-performance and self-perspective lead to a more accurate and vivid memory recall in simple action tasks, it remains an open question whether agency and perspective render the memory of a truly experienced episode susceptible to later modification. The purpose of the present functional magnetic resonance imaging (fMRI) study was to identify the modulational effects of agency and perspective on brain activity during memory retrieval and subsequent memory performance. Participants encoded a set of episodes by either imitating or only observing videos of short toy-based actions. During fMRI a week later, we cued episodic memory retrieval by presenting the original videos, or slightly modified versions thereof, either from 1pp or from 3pp. While we found decreased left lateralized hippocampal activity for previously only observed vs. self-performed actions, agency did not modulate memory performance. In general, subjects often misclassified modified videos as showing episodes experienced during the encoding session, when during fMRI, subjects had been presented (I) modified versions or (II) original versions from 3pp. Together, modified episodic cues can impair memory performance as well as re-experiencing encoded episodes from 3pp. Regarding different agentive states, when subjects miss subtle differences between encoded and cued episodes, the brain remains sensitive for detecting this mismatching information in order to update internal models.
9.45am
The formation and consolidation of spatial memories in children and adults
Patrizia Maier, Iryna Schommartz, Carsten Finke, Christoph J. Ploner, & Yee Lee Shing
Spatial memories like episodic memories rely on spatiotemporal associations and the hippocampus as a hub region. Spatial representations can be categorized as self-centred (egocentric) versus object-centred (allocentric). Egocentric representations involve episodic memories e.g., when remembering a previously travelled path as a sequence of left and right turns. Allocentric representations are episodic in the sense that they may involve detailed object-to-object-relations. The long-term consolidation of episodic and spatial representations is accelerated by prior knowledge in the form of schemas. Personal life experience defines the schema strength and may impact memory encoding, consolidation, and retrieval. In our study, we assess the formation and consolidation of spatial memories over a two-week period in three age groups. Our sample consists of 6- to 7-year-old children, 9- to 10-year-old children and 18- to 35-year-old adults. We use a behavioural spatial navigation task, where participants remember goal locations in a virtual maze, and we probe their memories in egocentric and allocentric conditions. Our preliminary results show better performance with increasing age in egocentric and allocentric spatial learning. Regarding consolidation, children in both age groups forget relatively more than adults over the two-week period. Overall, older study participants showed a superior ability to form and consolidate spatial memories. This may be due advanced brain maturation, more prior experience with spatial navigation situations, and availability of stronger pre-existing spatial schemas to build upon and consolidate faster.
Learning
Learning
Session chaired by Karoline Bading
10.30am
Why bother? Relational evaluative conditioning is moderated by the perceived probability of US occurrence
Karoline Bading & Klaus Rothermund
In Heycke and Gawronski (2020), pharmaceutical products (CSs) preventing positive or negative health states (USs) produced weaker evaluative learning than pharmaceutical products causing these health states. Such asymmetrical relational evaluative conditioning (EC) effects were found for evaluative ratings as well as for evaluative CS classifications, and an MPT analysis of the latter revealed an unqualified influence of US valence on evaluative behavior as measured by the CS classification task. While the weaker relational EC effects for "preventing" CSs are indeed well explained by a (non-propositional) influence of US valence on CS evaluation, the aforementioned pattern can also be accounted for by weaker propositional learning for "preventing" than for "causing" CSs. In the present study, we tested the idea that propositional learning for "preventing" CSs in Heycke and Gawronski (2020) may have been selectively disadvantaged by the use of rather extreme health states (as USs) that are highly unlikely for the typical participant. In a close replication of Heycke and Gawronski (2020), we found asymmetrical relational EC effects and an unqualified influence of US valence in the MPT analysis when using subjectively improbable health states as USs. By contrast, we found relational EC effects of comparable strength for "causing" and "preventing" CSs and no influence of US valence in the MPT analysis when subjectively probable health states served as unconditioned stimuli. The theoretical and methodological implications of these findings are discussed.
10.45am
Do we know what's in and what's out? The mental representation behind inward and outward articulation dynamics
Moritz Ingendahl, Ira Theresa Maschmann, Nina Embs, Amelie Maulbetsch, Tobias Vogel, & Michaela Wänke
People prefer linguistic stimuli with an inward (e.g., BODIKA) over those with an outward articulation dynamic (e.g., KODIBA), a phenomenon referred to as articulatory in-out effect. Despite its robustness across languages and contexts, the cognitive processes behind this effect are still at debate ‐ up to the discussion whether the effect is about inward and outward articulation dynamics at all. In the present research, we examine whether people have a mental representation of inward and outward articulation dynamics. In three experiments (\(N = 420\), two experiments preregistered) with an evaluative conditioning procedure, we systematically paired words with inward versus outward dynamics with positive versus negative valence. Although this evaluative conditioning procedure attenuated or even reversed the in-out effect, this moderation only occurred for words with the same consonant sequences as the conditioned words. For words with inward/outward dynamics but different consonant sequences than the conditioned ones, no moderation of the in-out effect was found. These findings imply that people do not have a generalized mental representation of inward/outward dynamics but instead evaluate stimuli based on stimulus-specific properties.
11.00am
There's more to life than "good" or "bad": Non-evaluative semantic stimulus relations in the evaluative decision task, revisited
Manuel Becker, Marie A. Jakob, & Karl Christoph Klauer
A recurrent debate centers on the malleability of the Evaluative Decision Task (EDT; Fazio et al., 1986) to the influence of non-evaluative semantic stimulus relations: There is little empirical agreement on whether and how evaluative and non-evaluative semantic stimulus relations (interact to) produce evaluative priming effects, and there is not even agreement on whether evaluative priming effects emerge in the absence of non-evaluative stimulus relations. An absence of priming effects in conditions where stimuli are only evaluatively related would be surprising inasmuch as it is hardly reconcilable with the dominant theories explaining evaluative priming effects via response activation. Resolving these conflicting results would thus inform theorizing on the mechanisms behind evaluative priming, and it would also have implications for sound measurement in the EDT, since in many applied studies, stimuli do not only vary on the evaluative dimension. In a first study (\(N = 81\)), we suggest that research on this topic so far has suffered from a confound that renders the EDT's measurement outcome difficult to interpret. In a second study (\(N = 81\)), we show that in a research design that does not suffer from this confound, evaluative and non-evaluative semantic stimulus relations co-determine the EDT's measurement outcome, and evaluative priming effects also emerge in the (relative) absence of non-evaluative semantic stimulus relations. We will also report results from an ongoing additional study in which only words from the semantically unrelated conditions are presented to participants.
The theory of affective meaning as proposed by Osgood et al. (1957) posits that the affective ecology of humans is shaped by three major dimensions—evaluation, potency, and activity—also known as EPA dimensions. Previous research has demonstrated that these dimensions largely account for a broad spectrum of affective experiences and has accumulated evidence for the robustness of the impact of the EPA dimensions for learning of affective meaning. However, so far previous research mostly concentrated on the investigation of a single or a subset of dimensions, did not disentangle confounded learning procedures, and only relied on the usage of restricted orthodox designs that do not consider the skewed distribution of the EPA dimensions in affective space. I present work in which I addressed these issues (Richter & Hütter, 2021). First, I employed a Brunswikian sampling procedure to select stimuli that allow for the implementation of a representative design, which acknowledges the natural relationships of the EPA dimensions in affective space. Second, I report two experiments that show how basic learning procedures—stimulus pairing versus stimulus exposure—differentially instigate the learning of affective meaning. Stimulus pairing elicited the learning of evaluation and activity, whereas stimulus exposure elicited the learning of potency and activity. I embed these findings in contemporary emotion theories, discuss implications for attitude learning, and reflect on epistemic and meta-theoretical aspects of this research. Finally, I argue that representative design enriches our methodological toolbox and inspires innovative theoretical accounts and the deduction of novel hypotheses that support a psychological science that is original and creative.
11.30am
Temporal statistical learning in the arcade game Whac-A-Mole
Nadine Schlichting, Josh Manu Salet, & Eckart Zimmermann
Statistical learning (SL) is the ability to automatically learn regularities of our world in order to predict future events. SL not only allows us to predict what will happen, but also when. Such temporal SL has received little attention thus far. We developed a paradigm inspired by the arcade game Whac-A-Mole (WAM) to study temporal SL. In WAM, participants score points by 'whacking' sudden-onset targets ('moles') arranged triangularly on a screen with the mouse cursor. Unbeknownst to participants, one target appears regularly (every 3 s) while the others appear at random. We found that although participants were unaware of the regularity, they unintentionally utilized this information and responded faster and more accurately to regular targets. As the regular location and action were directly coupled, participants could have learned to orient their attention to the regular location, prepare the regular action, or a combination of the two. In preregistered follow-up experiments (https://osf.io/sgcun/ and https://osf.io/3b78x) we assessed what exactly was learned to improve behavior. In one experiment we decoupled location and action by using a color response-mapping (pressing colored buttons associated to colored targets). We found no temporal SL of either dimension, suggesting that their coupling may have a superadditive effect. Indeed, when coupling action and location for the color response-mapping, temporal SL reoccurs. Additionally, when controlling for Simon interference by arranging targets vertically instead of triangularly, we find advantages of the regular target reflected in higher hit rates for regular action, but not for regular location. Taken together, these results indicate that timing is integrated in action; without it, temporal SL does not emerge.
11.45am
Origami folding: Hierarchical sequence learning of first step and inter-steps with practice
Fang Zhao & Robert Gaschler
The acquisition of sequential skills through practice is a fundamental ability of human activities, otherwise even simple tasks such as getting dressed would require a large amount of attention each time. Most studies on sequence learning used simple stimuli, such as pressing certain keys (e.g., Schumacher & Schwarb, 2009). However, a more fine-grained approach, targeting how the sequences are learned in a hierarchical sequence task, is lacking so far. We used an Origami folding task as the sequential learning task. Participants (\(N = 87\)) folded one out of five figures in four trials (i.e., each trial is a repetition of folding a figure). The results showed that there were significant improvements from trial 1 to trial 4 among all figures. Interestingly, the processing time for the first step was much longer than the inter-steps among all five figures, which may be due to the pre-programming of the sequences. Fluctuation in the inter-steps decreased with increased practice. As trial 4 was significantly faster than trial 3, sequence learning had not yet reached automation (ceiling effect). Further studies should thus test the influence of more additional practice sessions on sequence acquisition of a hierarchical sequence task.
Educational psychology
Educational psychology
Session chaired by Demian Scherer
4.00pm
More than the sum of its parts: A whole-picture advantage for learning with text-picture combinations
Demian Scherer & Stephan Dutke
People learn more deeply from text plus pictures than from text alone (Mayer, 2005; Schnotz, 2005). This multimedia effect was observed in numerous studies, with transfer measures often showing larger effects than knowledge measures. This pattern of results may arise because complete pictures showing arrangements of multiple smaller pieces provide an overall overview on the learning content: They provide information about relations between single parts and thus reveal structural features of the domain. If structural information contributes to the positive effect of pictures on learning from text, overall pictures should lead to better learning than partial pictures that show all relevant parts but not all relations between them. To test this hypothesis, in our experiment, participants learned about the geography of a fictitious kingdom either form text alone, from text passages plus partial pictures, or from text passages plus one overall picture showing the whole area of the kingdom. We measured learning performance with knowledge items (addressing content directly represented in the text or in the partial pictures) and transfer items (addressing overall relations that can also be inferred from combinations of partial pictures or text passages). Results demonstrate better learning from text and partial pictures than from text alone. Further, the overall pictures led to better performance than partial pictures for knowledge items and even more so for transfer items. We conclude that overall pictures contribute to multimedia effects especially when transfer is required: With an overall picture, less transfer steps are needed. Further, complete pictures also help to maintain smaller pieces of information by embedding them into an overarching arrangement.
4.15pm
Generating self-explanations: Self-generating new inferences, not paraphrases, boosts learning
Benjamin Harders & Mirjam Ebersbach
The present study investigated whether the positive effect of generating self-explanations on learning can be attributed to the generation of new relations between given facts or also emerges when paraphrases are generated by the learners. In a laboratory experiment, 118 participants generated self-explanations either on the basis of a text (a) that already contained relations between individual facts (complex condition) or (b) that presented the same facts in an unrelated manner (simple condition). There was no difference between the two conditions regarding the number of generated self-explanations in the learning phase. In a memory test after one week, participants were asked to judge facts referring to the original text as being true or false. Participants who had read the simple text performed better in the test than participants who had read the complex text (\(d = .32\)). The number and type of self-explanations and their relation to the performance in the test were analyzed. The number of self-explanations generated in the learning phase predicted the performance in the test only in the complex condition, but not in the simple condition. In addition, the number of paraphrases did not predict test performance. Thus, the self-explanation effect in the complex condition can be attributed to self-generation only, not to paraphrasing. Potential explanations for the finding that the self-explanation effect was absent in the simple condition, even though participants in this condition outperformed participants in the complex condition regarding learning success, are discussed.
4.30pm
Using group work to reduce procrastination among students
Markus Koppenborg & Katrin Klingsieck
Theorizing on procrastination as well as programs to reduce it mostly focus on person-related variables and neglect situational factors, such as group work. Findings from Social Psychology and Education Research show that group as compared to individual work can increase group member's motivation and performance, if the member perceives the own contribution as indispensable for the group product. A field experiment was used to test whether this type of group work can also lead to lower procrastination of an academic task and higher task performance as compared to individual work; and to explore positive and negative task-related affect. Student participants with a high level of trait procrastination (\(N = 58\), 37 female) were invited to complete an academic task over the course of ten days for a reward of €15. Instructions informed participants that they would work on the task alone (individual work condition) or as part of the group with their task representing an indispensable contribution to the group product (group work condition). We assessed procrastination of the task (using the Academic State Procrastination Inventory; Helmke & Schrader, 2000; \(\alpha = .92\)), whether the task was successfully completed, as well as the number of errors in the task, and positive and negative affect (using the Positive and Negative Affect Schedule; Krohne et al., 1996; \(.81 < \alpha < .88\)). Compared to individual work, task procrastination was lower, \(t(56) = 2.16\), \(p = .017\), \(d = 0.59\), and the number of completed tasks was higher in group work, \(\chi^2 = 5.09\), \(p = .022\), \(d = 0.62\). No significant differences were found regarding errors and positive and negative affect (\(p < .05\)). Results are discussed in the light of an innovative approach that draws on situational factors to prevent procrastination.
4.45pm
Enthusiasm, flow and the voice ‐ what happens in a magical number three learning condition?
Julia Bastian
How can enthusiasm in a learning environment be reached? Is an enthusiastic voice more effective to activate learning than a neutral or depressed voice? University students heard either an enthusiastic, a depressed or a neutral voice reading learning material. Participants were assigned to one cell of a three (enthusiasm: enthusiastic vs. neutral vs. depressive) factorial between-subjects design. Low and high mental load and effort in the working memory of the learner, a low and high extent of flow experience and the mood before and after the voice induction were explored according to the learning condition. Liew et al. (2015, 2017) and Beege et al. (2020) identified mental or cognitive load as a central moderator between enthusiasm and learning outcomes. According to their perspective emotions inevitably induce mental load. In accordance with the view of emotional response theory enthusiastic verbal and nonverbal cues of an instructor or teacher can induce higher positive emotional states in learners, which in turn, enhance affective perceptions, intrinsic motivation, cognitive outcome and enact learners' approach behaviours in the learning process. Contrary, in the view of cognitive load theory enthusiasm may increase extraneous cognitive load as an additional processing in the mind, which negatively impact emotion, affective perceptions, intrinsic motivation, and cognitive outcome. Additional insights into the boundary conditions of the effects of emotions on learning will be provided. Results, implications and suggestions for future research are discussed.
5.00pm
Experimental evidence for feedback inflation and the positive feedback bias: Judgments of vs. feedback to students with German and Turkish names
Anna K. Nishen & Ursula Kessels
Feedback is a central tool for learners to understand how they can improve (Hattie, 1999) and it therefore needs to be accurate. However, feedback is also an inherently interpersonal situation (Müller & Ditton, 2014). Compared to an initial non-communicated judgment, teachers may therefore adjust their feedback to make it more palatable and motivating. However, the difference between non-communicated judgments and feedback may be greater for students belonging to negatively stereotyped groups. Prior research has established a persistent negative bias towards ethnic minority students in non-communicated judgments (e.g., Sprietsma, 2013) and a positive bias in feedback thought to be elicited by the concern that criticism may reflect prejudice (e.g., Harber et al., 2010, 2012). We tested these hypotheses using a 2 (migration background: Turkish vs. none; within subjects) \(\times\) 2 (evaluation condition: judgment vs. feedback; between subjects) design in a sample of 132 German teacher students. After being randomly assigned to the judgment/feedback condition, participants each judged/gave feedback on two essays supposedly written by a student with a Turkish vs. German name. A pilot (\(n = 20\) teacher students) had shown that both essays were of the same low quality. Mixed ANOVAs showed that participants in the feedback condition gave better grades, were more positive in their ratings, and included more positive comments in their open text than participants who simply judged the essays. Moreover, participants were more positive in their impressions of the student with a Turkish compared to the student with a German name—unexpectedly, this was the case in both judgments and feedback. We discuss implications of these findings for students, e.g., their self-assessment and motivation.
5.15pm
Emotional reactivity to music: A mediating factor for developing musical abilities
Daniel Fiedler, Tatiana Rödel, & Anne Frenzel
Researchers widely agree that the development of musical abilities is a very complex process involving many different factors. In the present contribution, we propose to consider emotional reactivity to music as one of them. While the strong interconnectedness between music and emotions is widely accepted, intriguingly, they have not been taken into account yet in the context of musical ability development. Our study aims, based on ideas of musical ability development and theories about music and emotions, to analyze the relationships between emotional reactivity to music with musical engagement and musical perception abilities. The sample consisted of 127 adults (77.2% female, age: \(M = 25.12\) (\(\mathit{SD} = 7.02\))), and the procedure involved participants listening to 15 relatively unknown soundtrack excerpts. Prior to the music excerpt exposure, participants' musical engagement was obtained through self-report. Afterwards, emotional reactivity to music was operationalized during the music excerpt exposure through self-reports on arousal and valence as well as skin conductance response and heart rate. Finally, participants' musical perception abilities were assessed using an adaptive test battery. The findings showed that the measures of emotional reactivity to music were significantly correlated with the indexes of musical engagement such as music listening or musical training and practice (\(r\) between .18 and .30, all \(p \leq .05\)), which in turn were significantly correlated with musical perception abilities (\(r\) between .19 and .59, all \(p \leq .05\)); but musical perception abilities and emotional reactivity to music were uncorrelated. We conclude that emotional reactivity to music is relevant for engaging in musical activities and thus a potentially important precursor of musical ability development.
Pear Odeum
Experimental engineering psychology and human factors
SymposiumExperimental engineering psychology and human factors
Symposium organized by Stefan Brandenburg and Martin Baumann
The facilitated integration of technology into people's lives highlights the importance of examining its impact on experience and behavior. Experimental approaches help to determine the underlying psychological processes of this impact. This symposium summarizes experimental studies examining various contexts of technology use and psychological aspects of Engineering Psychology and Human Factors. The first talk of Yuxuan Guo investigates whether people shift back and forth from ambient to focal processing based on internal triggers, such as the shift within and across the processing of tasks while no external event is given. The second talk of Matti Krueger addresses how technology can be designed to augment situation awareness, specifically in the context of traffic. Several prototypes and associated user studies are presented to demonstrate the utility of this approach. The third talk of Agnes Rosner describes an experimental study addressing memory retrieval processes in comprehension of dynamic traffic situations by using the so called looking-at-nothing phenomenon. The fourth talk of Tim Schrills describes a study in the field of Human-AI Interaction in Personalized Medicine. Here, the authors ran an experiment with diabetic patients assessing their subjective information processing awareness. The fifth talk of Franziska Babel addresses a specific aspect of human-robot interaction: robot bullying. In an online experiment it was investigated which interaction patterns varying from friendly to severely bullying participants preferred when interacting with a robot in a goal conflict situation. The sixth talk of Thilo Kremer examines the impact of four contexts of technology use like online shopping on the importance ratings of 12 standardized user values.
8.30am
Eye movement patterns in complex tasks: Characteristics of ambient and focal processing
Yuxuan Guo, Jens R. Helmert, & Sebastian Pannasch
Analyzing the time course of eye movements during scene viewing indicates that people progress through two distinct modes of visual processing: an ambient mode, serving global spatial orientation in the visual environment, followed by a focal mode, facilitating a more elaborated and object-centered analysis. However, the shifts between ambient and focal processing modes have mainly been identified relative to environmental changes like the onset of various visual stimuli, scene cuts, and subjective event boundaries in dynamic stimuli. The results so far do not allow conclusions about the nature of ambient to focal processing beyond the influence of externally triggered events. It remains unclear whether people shift back and forth from ambient to focal processing also based on internal triggers, such as the shift within and across the processing of tasks without external events. In our study, ambient to focal attention was investigated in an active task solving paradigm. The used Rubik's Cube task is a multi-step goal-driven task, which can be broken down into smaller sub-tasks that are performed serially. The time course of eye movements was analyzed across different levels of task processing with and without visual changes in the task environment. The results suggest that initial ambient gaze patterns are followed by a switch to more focal viewing across many levels of processing. More importantly, the ambient to focal processing works as the basic attentional mechanism in task processing, which does not seem to be influenced by changes in task performance. The present findings suggest that ambient and focal eye movement characteristics might serve as a probe for the attentional state in task processing. Future work should therefore focus on more applied settings.
Safe mobility in traffic environments heavily rely on an individual's ability to perceive, comprehend, and anticipate relevant information for collision avoidance. Collectively termed "situation awareness", the lack thereof continues to be a major contribution to accidents attributed to human errors. How can we design technology to mitigate this? This presentation begins by addressing the critical factors for awareness formation, particularly in terms of those that can benefit from machine assistance. We argue in favor of a class of human-machine interaction that is best labelled 'augmentation', so termed as it requires active user engagement, preserved awareness, and minimal interaction complexity. This implicitly places authority (and liability) on the human and not the machine. A basis for this is established by fundamental theories and established findings on human sensorimotor processes. Finally, we demonstrate how this approach can lead to prototypes for augmenting situation awareness in driving, with results from user studies.
9.00am
"Looking at nothing" while comprehending a dynamic driving scene
Agnes Rosner, Wiebke Frank, Kristin Mühl, & Martin Baumann
While driving, people build a visuo-spatial representation of the environment and retrieve pieces of information from it during situation comprehension, like actual speed limits that are often no longer visible to the driver. While retrieval is crucial for safe driving behavior, memory processes have not been well investigated, possibly because suitable methods were missing that would allow inferences about retrieval processes during situation comprehension. We tested if eye movements based on the so-called "looking-at-nothing"-behavior (LAN) can be studied to gain insight into the underlying retrieval dynamics. Previous research using static scenes has shown that LAN can reveal which information is retrieved from memory as people tend to fixate spatial positions of visually encoded information, even if it is no longer visible at that location. In one experiment, participants (\(N = 33\)) viewed dynamic traffic scenarios on a reduced screen and evaluated auditory statements relating to different levels of situation awareness (SA) while eye movements were recorded. Results showed that when retrieving information of SA levels 1 and 2, subjects fixated more on emptied spatial locations associated with information relevant for the probed SA level. The retrieval of anticipations (SA level 3) resulted in more sequential gaze transitions that corresponded to the spatio-temporal dynamics of future driving behavior. The results support the idea that drivers build a visuo-spatial representation of the driving situation. Different gaze patterns when retrieving level-specific information indicate divergent retrieval processes. Thus, LAN can be used to study memory retrieval to advance knowledge on driving-related situation comprehension.
9.15am
Subjective information processing awareness in hybrid closed loop systems ‐ human-AI interaction in personalized medicine
Tim Schrills & Thomas Franke
While interacting with artificial intelligence (AI) in systems, users experience automated information processing, which can remain hidden to them. Grounded in research on human-automation interaction, we propose Subjective Information Processing Awareness (SIPA) as a metric of user experience in Human-AI interaction contexts. SIPA can be defined as the experience of being enabled by a system to 1) perceive, 2) understand and 3) predict its information processing. For example, in hybrid closed loop systems in the medical domain as for example in diabetes therapy users interact with AI systems on a daily basis already today. Recently, automated insulin-delivery (AID) systems calculate the insulin need based on physiological measures. However, they still need information provided by the user e.g., on physical activity. For a successful implementation of AID systems, it is crucial that users trust and understand the performed insulin need calculation. The objective of the present research was to advance understanding about the experience of people interacting with AI in the context of personalized medicine. In an experiment with diabetic patients, we modified the amount of information an AID system presents to the user as a between-groups variable. To this end, participants were repeatedly shown the system's calculated insulin need as well as the input information it processed. We also examined, to what extent participants were able to predict the calculated insulin need. The analyses will focus on 1) the development of the dependent variables after repeated interactions, 2) the relationship between SIPA performance and 3) the correlation between trust and SIPA. We expect implications regarding the effects of explanations on subjective experience in Human-AI interaction.
9.30am
An interactive online study to investigate the causes of robot bullying
Franziska Babel, Philipp Hock, Johannes Kraus, & Martin Baumann
When service robots enter public spaces the so-called 'robot bullying' can occur: the robot is impeded in its task execution or gets damaged. To prevent robot bullying, human motives for these disruptive actions need to be understood. In the interactive online study, the 41 subjects were told that they would meet a cleaning robot in the hallway and had the choice of interacting with the robot or walking by. If they chose to interact, they could choose one of three categories (friendly, mild bullying or severe bullying) and then a specific behaviour (e.g. touching) for which they saw the consequence afterwards (e.g. robot is damaged). This was repeated until the participant decided to pass the robot finally. The bullying categories were based on a pilot study with 19 subjects who rated the videos regarding the bullying severity. The majority (82%) chose not to interact with the robot. Those who interacted did so in a friendly manner. Only mild abuse and no severe bullying was selected. When asked about the reasons for their behaviour, 27% said they did not want to disturb the working robot, and 20% said they were not interested in the robot. Lack of trust in the robot predicted disinterest. The results might indicate a more benevolent behaviour towards robots in public than expected, given that in an online study, one could have anonymously mistreated the robot. This online study contributes to the understanding of robot bullying and provides a methodological contribution to investigate robot bullying interactively online.
9.45am
The effects of context on the importance of user values in human-computer-interaction
Thilo Kremer & Stefan Brandenburg
Considering user values in software development has a variety of positive effects. Values motivate users to use software, increase customer loyalty and the social acceptance of technologies. Theoretical approaches like Value Sensitive Design (VSD) propose 12 user values for software development. We conducted an online experiment with 128 participants that aimed to answer three research questions. First, which user values should be considered in software applications (absolute relevance)? Second, how do users prioritize these values (relative relevance)? Third, are users willing to accept additional efforts to support values in software development (cost acceptance)? Participants were randomly assigned to one of four conditions (IV 1, between-subjects) email program, social network, search engine, online marketplace reading 12 vignettes (IV 2, within-subjects), each capturing one of the VSD-values in randomized order. Four items for each vignette assessed the absolute relevance of values on a 7-point Likert Scale ranging from 1 "do not agree at all" to 7 "completely agree", and two items assessed their willingness to accept costs. Finally, participants ranked the 12 values concerning their relative importance. Results revealed significant differences in the absolute relevance of user values, with higher relevance of "privacy", "trust", and "sustainability". "Courtesy" and "calmness" were evaluated as comparatively less relevant. The ranking task confirmed these results. We also obtained significant differences in the importance of values between conditions. Finally, participants were willing to accept costs for considering their values in software products. Based on these findings, we conclude that user values should be considered in software applications.
Human factors
Human factors
Session chaired by Aike Horstmann
10.30am
Automatic pain monitoring through facial expressions: A deep learning approach
10.45am
The occurrence of miscues in decision support systems: Effects on detection performance of airport security screeners
David Huegli, Alain Chavaillaz, Juergen Sauer, & Schwaninger Adrian
Decision support systems (DSS) are one type of automation that supports human information acquisition to improve decision-making. At airport security checkpoints, the use of explosives detection systems for cabin baggage (EDSCB) as a DSS support airport security screeners by highlighting areas in X-ray images that might contain explosive material. However, EDSCB are not perfectly reliable . They sometimes fail to give the correct response, resulting in three types of failures: misses, false alarms or miscues. When the EDSCB displays a miscue, screeners experience an explosives false alarm but another prohibited item (e.g. a gun or knife) is located elsewhere in the bag. The screeners potentially miss the other prohibited item, because they were too focused on the cue provided by the EDSCB. Misses and false alarms of DSS are well investigated, but hardly any studies addressed the effect of miscues on detection performance. The present study aims to examine the effects of each different failure types. We tested 115 professional airport security screeners with realistic X-ray images of cabin baggage. They were randomly assigned to three different experimental conditions: Miscue-prone, false alarm prone or multiple failures. Screeners had to detect bombs, guns and knives and were supported by EDSCB. Results show that screeners missed more knives when EDSCB miscues occurred compared to when EDSCB did not issue an alarm. Also, miscues misled screeners in thinking that X-ray images contain bombs while missing other prohibited items. We conclude that miscues of DSS have a negative impact on detection performance during a visual inspection task and that miscues are a problem to be considered when working with EDSCB at airport security checkpoints.
11.00am
Field study regarding the work duration for the visual inspection of X-ray images of passenger baggage
Daniela Buser, Yanik Sterchi, & Adrian Schwaninger
At airport security checkpoints, security officers (screeners) inspect X-ray images of passenger baggage for prohibited items. Because EU regulation limits continuous inspection to 20 minutes, screeners typically thereafter change to another task at the checkpoint. With a new technology, remote screening, the task of X-ray image inspection is conducted in a separate room away from the checkpoint. This makes it operationally challenging to maintain the task rotation at 20-minute intervals. The question arises whether screeners can inspect X-ray images for longer periods of time and how this affects their performance and well-being. In a field study at a European airport, two groups of screeners inspected X-ray images of cabin baggage over a four-month period. One group analyzed X-rays for 20 minutes and then moved to a different work position (control group), while the other group analyzed X-rays for up to 60 minutes (experimental group). A group comparison showed that screeners did not differ on various performance measures (hit rate, reject rate, or processing time). The examination of the screening sessions conducted by the experimental group revealed that when the workload was low or average, the screeners were able to maintain their performance for up to 60 minutes. However, if the workload was high, the hit rate dropped significantly. In interviews and questionnaires, screeners reported that they usually had difficulties in maintaining attention after 30 to 40 minutes. Our results imply that longer screening sessions between 30 to 40 minutes could be tested and introduced. Thereby it is advisable to monitor and set a minimum hit rate that should be achieved during these prolonged screening sessions.
11.15am
Development of a textile-integrated sensor system to assist the rehabilitation after cruciate ligament injuries: Comparison of user experience and usability evaluation of initial prototypes
Aike Horstmann, Barbara Hemsen, Jarek Krajewski, & Sebastian Schnieder
The yearly incidence of cruciate ligament ruptures in Germany is estimated to be 45 per 100,000 residents. Even after successful surgical treatment, a great number of patients experience instability in the injured knee (giving-way phenomenon), resulting in avoidance behavior, uncertainty regarding the own physical resilience, and functional limitations in everyday situations. To address this problem, an innovative system for reducing critical movements and subsequently improving the rehabilitation after cruciate ligament injuries is being developed. Motion sensors are integrated in a smart knee bandage to detect critical movements such as giving-way events. Via vibratory feedback the patient will be notified and enabled to adjust the knee position accordingly. The present study aims at comparing usability and user experience aspects of three initial prototypes (P) of the described knee bandage. The prototypes mainly differ in how the sensor and battery elements are fixated. P1 (n = 6) and P2 (n = 5) have no sensor casings, one big battery case, and all elements attached directly to the textile. In contrast to P1, P2 has sticked electric lines. P3 (*n* = 5) has sticked electric lines, sensor casings, and two smaller battery cases. Furthermore, the sensor elements are fixed on a separate mesh substrate which is added after the patient put on the textile base. A one-way ANOVA revealed no significant difference between the three prototypes regarding haptics, usefulness, originality, and wearability of the knee bandage. This is most likely due to a ceiling effect since all prototypes were evaluated strikingly positively. The content of the free text inputs by the testers will be evaluated to gain insights for the further development.
11.30am
Evaluation of simulated recognition aids for human sensemaking in applied surveillance scenarios
Mary E. Frame, Robert Patterson, Anna Maresca, & Amanda Christensen
In surveillance operations, accurate mental models must be constructed to make effective decisions, which requires making sense of dynamic information as it unfolds. For intelligence analysts, sensemaking is challenging because it frequently involves making inferences about uncertain data. One way to enhance sensemaking may involve collaboration from a machine agent such as Project Maven, an established algorithm that directs analysts' attention to people and vehicles in scenes. We simulated actions of a Project Maven as well as that of a machine agent that directed attention to EEIs. Sensemaking was assessed by measuring participants' ability to predict events and identify signs. We found that simulated agents directing participants' attention to EEIs improved EEI identification but directing attention to people and vehicles (emulating Project Maven) did not. Participants' higher-order sensemaking was not enhanced by either type of machine agent. This demonstrates that guiding attention to features in a scene improves identification whereas indiscriminate steering of attention to entities in the scene does not improve understanding the holistic meaning of events. These results contribute to our goal of determining which human-machine systems improve the sensemaking capability of intelligence analysts in the field.
11.45am
Effects of color, route-instructions and birdseye's map in human wayfinding
Manish Kumar Asthana
Human beings are living in a dynamic and complex environment. They are often exposed to extensive wayfinding information, which might result in ineffective and poor decision-making. This study examines the impact of color, route instruction, and Birdseye's map in human Wayfinding. An experiment designed with four different conditions as follows: (i) Instruction-Bird's eye (IBV condition), (ii) Instruction-No-Bird's eye (INBV condition), (iii) No-Instruction-Bird's eye (NIBV condition), and (iv) No-Instruction-No-Bird's eye (NINBV condition). Participants' task was to find their way to a goal in a virtual environment with either a colored or grey ground. Fifty-four healthy college students (\(18-35\) years) took part in the study. All participants were requested to complete informed consent and a set of screening questions, and a demographic form. Participants were given a training section (exploration phase) to explore and learn about the environment and system. The experiment was divided into two halves, with a 5-minute break to reduce the trial learning effect. In the first section, individuals were randomly assigned to one of four colored or grey randomized trials. Participants finish the second half of the experiment after taking an activity break. It was found that color has a significant effect on landmark identification and wayfinding performances in external environments. The results revealed that both route instructions and Birdseye's map helped individuals in Wayfinding. Specifically, trials 1, 2, and 4 showed a significant effect of color, leaving the grey condition non-significant. In general, individuals' landmark identification and wayfinding performance seem to have benefitted from color, route instruction, and Birdseye's map in the virtual environment.
Interaction with artificial intelligence – An interdisciplinary perspective
SymposiumInteraction with artificial intelligence – An interdisciplinary perspective
Symposium organized by Nadia Said and Andreea-Elena Potinteu
Artificial intelligence (AI) applications are increasingly used in everyday life. Whereas some of them are widely accepted (e.g., automatically compiled playlists), others are highly controversial (e.g., use of AI in the classroom). While on the one hand, the usage of autonomous cars could save about 50 million lives within the next 50 years this AI application is still viewed with skepticism. On the other hand, people already interact with AI daily online; often unaware of how this interaction might influence their beliefs and behaviors. This shows that the question about what kind of factors influence people's willingness to interact with AI is of utmost importance. Aim of this symposium is to present novel insights into human-AI interaction by bringing together research from cognitive psychology and philosophy to address two questions: (i) what factors are influencing our perceptions of and ultimately our interaction with AI and (ii) how does interacting with AI in turn influences us? To provide answers for those questions the first talk focuses on the role of AI risk-opportunity perceptions as well as AI knowledge in people's willingness to use AI-based applications in different fields. Going beyond abstract algorithms, the second talk then tackles the question under which conditions people are willing to help robots. The third talk focuses on the cognitive factors that are relevant for cooperation with AI. The fourth talk will then give an overview on how our interaction with AI in turn shapes and challenges people's perception of social interaction. In the final talk, the question how AI should be modeled as interaction partners is discussed. Jointly, these talks will give a broad overview on human-AI interaction by examining the topic from different perspectives.
4.00pm
Knowledge about artificial intelligence (AI) moderates the relationship between risk and opportunity perception of AI and people's willingness to use AI-based applications
Nadia Said, Rebekka Schwesig, Irina Brich, Jürgen Buder, & Markus Huff
Surveys around the world show that the public perceives artificial intelligence (AI) as a two-edged sword: as a risk but also as an opportunity. So far however, there has been no research investigating how people's risk-opportunity perception relates to people's willingness to use AI-based applications. In the presented study we investigated the relationship between (i) risk-opportunity perception of AI, (ii) AI knowledge, and (iii) willingness to use AI. To this end, we conducted an online study with \(N = 245\) participants. For the risk-opportunity assessment participants had to judge one out of four different scenarios describing AI applications in various contexts (medicine, transport, media, psychology). AI knowledge was assessed using a newly developed 12 item questionnaire. Willingness to use AI was measured by giving participants the option to choose between using vs. not using AI (for example to analyze a beforehand filled out health questionnaire) at the end of the study. As hypothesized, risk-opportunity perception of AI correlated positively with the probability to use AI. Furthermore, accurate knowledge moderated this relationship in that, for participants with higher AI knowledge the risk-opportunity assessment plays a negligible role regarding their willingness to use AI. By including opportunity perception as a predictor variable for willingness to use AI, rather than solely risk perception, this study expands existing research that is mainly focused on risk-taking behavior only. Moreover, results underline the importance of accurate AI knowledge as a moderating factor regarding the willingness to use AI applications.
4.15pm
People's reported willingness to help robots
Andreea-Elena Potinteu, Nadia Said, & Markus Huff
Society is increasingly using technology in everyday life, including robots. In turn, the chances of people meeting a robot in an unusual situation also increase. For example, would people be willing to help a robot that has trouble going over an elevated sidewalk? And what variables would contribute to people's willingness to help? To explore this matter, an online study aimed at \(N = 312\) is being prepared (pre-registered at https://aspredicted.org/12N_2XK). Participants read 9 different scenarios from 3 domains (delivery, service industry, medical) in which an encountered robot needs help. Each scenario appears two times, in random order, followed by one of two possible visual stimuli: a humanoid or a machine-like robot. Participants decide if they would help the robot in each scenario. Attitude towards robots, knowledge (using a new robot knowledge test), and anthropomorphism are measured. The latter variable is measured using the Waytz instrument (2019) and manipulated in the experimental run using the 2 types of stimuli mentioned above. We will analyze how attitude, knowledge, and anthropomorphism relate to people's helping behavior computed as the proportion of help responses and the response time. Based on literature, we expect a high knowledge score to lead to low helping scores. On the other hand, a high anthropomorphism score and positive attitude are expected to relate to a higher helping score. We expect the humanoid robot to evoke more help responses as the machine-like one. Understanding the variables that contribute to people's willingness to help robots is relevant for a successful integration of robots into our society. Plus, understanding people's decision about helping robots or not can shed further light on how people mentally categorize robots.
4.30pm
Cognitive functions for cooperation with artificial intelligence in a game setting
Fritz Becker, Jürgen Buder, & Markus Huff
A lot of work is currently done on creating agents that are more adaptive towards the behavior of their human partners. These agents are generally trained by playing thousands of episodes cooperatively with themselves. That leads to problems when they are faced with a partner that does not act in a way the agent can foresee, like a human. In a cooperative game, it is vital that both partners can learn and adapt to what the other does. In work on the adaptability of agents, the fact that humans adapt to the agent too is widely ignored for good reasons. But it is an interesting question nonetheless: How does a human adapt to the behavior of an agent in a cooperative game? We created an experimental paradigm in which the participants play a cooperative version of the famous game Tetris together with a bot. The collaborative element is added by giving one partner the responsibility to choose a piece and the other to place it according to the Tetris rules. In the first experiment, the bot selects the pieces. It follows strict rules and does not alter its strategy to harmonize better with the human. That allows the human participant to learn the rules of the bot implicitly throughout many rounds. We test this by letting the participants predict the action of the bot. We assume that if the participants understand the bot, they should be better at predicting its choices than a control group with a randomly acting bot as a partner. This hypothesis is confirmed in an experiment on \(N = 87\) participants. In this experiment, we also show that while the group that has the smart bot as a partner is learning to anticipate the bot's actions, they gain no explicit knowledge of the rules it follows. We will discuss, whether this process is similar to social learning processes.
4.45pm
Is it advisable to take the intentional stance towards AI-systems?
Tobias Schlicht & Eva Weber-Guskar
Technological progress in the development of AI systems that populate our private environment, not only the workplace, call for new theoretical questions, practical challenges and ethical questions to do with social interaction. For a long time now, we have not only been interacting with fellow humans (and non-human animals) but also with chat bots, via smart devices, avatars, even robots. What is the effect of such an increasing presence of AI-systems in our daily lives? Theoretically, can we simply apply our philosophical/psychological theories of human-human social interaction to human-machine social interaction? Or do we have to adjust our assembly of cognitive strategies to make sense of each other? Taking one step back, should we even take the intentional stance towards such robots or tactically retreat to the functional stance? Should we conceive of them as friends, colleagues, slaves, machines? And what are the consequences for the future design of such systems, i.e., even if technologically possible, should we design them such that it is more likely that they invoke the one or the other stance in us? What would be advantages and risks of taking either of these stances towards intelligent AI-systems? This talk will explore some of the dimensions of this topic.
How do we understand others? According to the person model theory of understanding others (Newen 2015, open mind.net), we rely on person models, situation models and a self-model. This enables us to account for 1. the social roles and the individuality of the other person (person models), 2. to register the relevant situation the person is in (situation model) and to relate both to relevant aspects of oneself (self-model). Furthermore, this enables us to develop a social understanding on three levels, namely 1) online understanding of others realized as coordinated interaction in a type of situation, 2) understanding of others as rule-followers (e.g. according to group models indicating typical social roles) and 3) understanding of others as having an individual mindset of attitudes (relying on person models of individuals). Can we use these characteristics to evaluate existing robots as interaction partners and can this help us to describe demands to improve artificial interaction partners? The aim of the talk is to deliver a positive answer. Most robots are shaped to be sensitive to only one type of signal about the human interaction partner. Then there is an endeavor for shaping robots in the direction of acting according to situation models. But there is still a big gap since robots so far do not include person models neither for groups of persons nor for individuals. Advanced social interaction needs to include such models because our social competence is intensely shaped by knowing with whom we are interacting. Then our behavior in the same situation is adjusted to the interaction partner. This is one central challenge for social robotics.
Azure Lounge
The active information sampler - Information sampling approaches to judgement and decision making
SymposiumThe active information sampler - Information sampling approaches to judgement and decision making
Symposium organized by Linda McCaughey
Cognitive-ecological approaches have emphasised the influence of the information sample on judgements and decisions. Yet, samples are often actively solicited, implying a causal path from the cognitive processing stage back to the sampling stage, adding rich endogenous sampling aspects to the framework. Niese will demonstrate how the sample available to the decision-maker can induce framing effects without assuming biased information processing. Prager will discuss how the evidence conveyed by a sample of character traits influences the decision to truncate an unfolding sample, resulting impression judgements, and perceived target homogeneity. Also concerned with sample truncation, McCaughey will present findings suggesting that the sampling process is influenced by cost considerations. And although decision makers are sensitive to changing information sampling costs, this sensitivity is limited. Alves will introduce a sampling principle according to which people often rely on samples of distinct and surprising information, among which negative information is overrepresented. A number of well-known evaluative biases can be explained by the priority of distinct information in the formation of attitudes. Biella will elaborate on how an individual's sampling goal influences the sources they select. In sampling social information, a hedonic sampling goal induces a stronger tendency to move on to the next source after negative information than an epistemic goal. Focussing on the interplay between information sampling and contingency inference, Bott will present evidence that despite premature exploitation, self-determined sampling can, under certain conditions, foster the proper use of the sampled information for contingency judgements.
8.30am
It's good and it's bad: Sampling processes can generate decision-framing effects
Zachary Niese & Mandy Hütter
Framing a choice in terms of gains versus losses can have a dramatic impact on peoples' decisions, sometimes completely reversing their choices. This decision-framing effect is often assumed to stem from individuals' inherent motivational biases to react more strongly to negative information. However, more recent work suggests these decision biases can also stem from biases in the information samples based on which people make their decisions. Here, we test how biases in the frequency of information people have about each decision option can produce decision-framing effects via a sampling process. Specifically, we hypothesize that a gain versus loss framing determines whether people probe their memories for positive or negative information about each decision option. This can lead to inaccuracy if there are biases in the amount of information people have about each option. That is, when people have more information about one option, it is possible that they retrieve both more positive and more negative information about it, creating a bias to select it as being both more and less likely to result in success, depending on decision framing. Three experiments show that people's decisions are more accurate with a gain (vs. loss) framing when a high (vs. low) frequency option has a higher proportion of success; but decisions are less accurate with a gain (vs. loss) framing when the high frequency option has a lower proportion of success. The current results suggest that decision-framing effects do not necessarily indicate a motivational bias. Instead, sampling processes may underlie decision-framing effects.
8.45am
Impression formation from self-truncated sampling on social groups
Johannes Prager & Klaus Fiedler
The condition that judging individuals themselves can decide whether to stop a sequentially unfolding information sample or to seek further evidence might constitute a sufficient condition to classical inter-group phenomena, namely out-group homogeneity and out-group polarization/derogation. Typically, social out-groups (i.e., groups we are not part of) are perceived to be more homogenous in themselves, as compared to in-groups (i.e., groups we are part of; Linville, Fischer, & Salovey, 1989; Quattrone & Jones, 1980). We create a simplified inter-group sampling context by providing participants with a sequence of traits, characterising each a member of the current target group. Participants can decide for themselves when to stop this sequence of information. We predicted and regularly observed that early-truncated samples result in higher perceived within-group homogeneity and more polarized impression judgements compared to samples that were further expanded. Sample truncation and judgements are, however, not exclusively determined by the sample input alone, but moderated by properties and instances of cognitive processing in the mind of the judging individual. We propose a series of diagnosticity factors impacting how the incoming information is weighted and considered. Extreme and negative valence, but also properties of multi-dimensional density regularly impact sample truncation (i.e., sample size), judgement extremity/strength and perceived group homogeneity.
9.00am
Adaptively trading off benefits and costs of information in sample-based decisions
Linda McCaughey, Johannes Prager, & Klaus Fiedler
Searching for information prior to a decision is one of the most common decision-related activities. After all, more information almost always leads to better decisions. However, it also comes at a cost, financial or otherwise. To examine how well people can trade off those costs and benefits, we conducted two experiments, in which a total of 158 participants made sample-based decisions involving costly sampling and reward and punishment for correct and incorrect decisions, respectively. Across four blocks, the ratio of information cost and payoff changed, either due to a change in only one of the constituting parameters for better relative comparability (Exp. 1), or due to changes in both parameters (Exp. 2). Overall, participants were sensitive to the cost ratio changes, meaning that they adapted their search behaviour to the changes, with regressions of sample size on cost ratio resulting in a mean coefficient of \(M_r = -.24\), \(t(72) = -10.08\), \(p < .001\), for Experiment 1, and, \(M_r = -.17\), \(t(83) = -7.03\), \(p < .001\), for Experiment 2, potentially reflecting the lower comparability between cost ratios in Experiment 2. This indicates a non-negligible but small degree of adaptivity. Interestingly, participants sampled both too much and too little information, depending on the specific cost ratio. This shows how important it is to include more than one set of task parameters when investigating information search in decision making. On the one hand, this allows for the examination of adaptation to change, an important aspect of the real world. On the other hand, it safeguards against premature and overly general conclusions.
People's impressions, attitudes, and judgments necessarily rely on samples of information. I introduce a sampling principle according to which people seek distinct information that is rare and diverse, and that allows to differentiate between contexts, objects, people, or groups. Among distinct information samples, however, negative information is overrepresented. This follows because in most information ecologies, negative compared to positive information is less frequent, but more diverse. Consequently, when perceivers sample distinct information, resulting impressions, attitudes, and judgments will be negatively biased.
9.30am
Navigating the social environment: A sampling approach to trustworthiness
Marco Biella & Mandy Hütter
The social environment is the collection of all possible social partners. Navigating such an environment is not easy. People have to interact with a social partners, collect information about the interaction, and use such information to form the impression guiding subsequent behaviors. The information search process shows specific sampling patterns depending from the goal at hand. Generally, people maximize their exposure to positive information when no explicit goal is available, truncating the sampling process early after negative information. Consequently, impressions will be based on unreliable negatively-skewed samples triggering small-sample bias. But information search strategies change with an epistemic goal. Here, the tendency to truncate sampling after negative information vanishes as early truncation interfere with the epistemic goal. Now information quantity is more rewarding than its valence and the rational behavior is to collect as much information as possible despite its content. In four experiments (\(N = 402\)), we demonstrated that early truncation, and small-sample bias, occur when no explicit goal is provided. This tendency is reduced under an epistemic goal. In a totally unconstrained information sampling paradigm, participants went through the sampling phase knowing that later they will be asked to perform a task whose completion requires the sampled information. Knowing the final task in advance allowed our participants to tailor their sampling behaviors to their goal. If the task implied an epistemic goal, participants avoided early sampling truncation and obtained larger samples. Final impressions were directly related to the available samples. If a bias is there, it is in the sampling process rather than in the information integration stage.
9.45am
On the interplay between information sampling and (pseudo)contingency learning in judgments and choices
Franziska Bott
Ideally, correlations between variables are inferred based on their co-occurrences, like the individual frequencies of choice options and outcomes co-occurring (joint frequencies). Thus, to determine contingencies (e.g., between options and outcomes), information needs to be sampled and integrated at the level of aggregation appropriate for the task at hand. A wide range of studies, however, showed (a) that simple information sampling strategies can lead to premature avoidance of choice options believed to result in negative outcomes, leading to skewed samples, and (b) that individuals may base their contingency inferences on more aggregate data than would be appropriate, namely on observed skewed marginal frequencies instead of the observed joint frequencies. As a consequence, subsequent choices were demonstrated to be sub-optimal. In this project, we investigated various factors that may affect information sampling and thus lead to asymmetries in the information available, which in turn might provide the basis for erroneous contingency judgments. But just as the characteristics of the available sample affect the kind and size of the contingency that is inferred, previously learned or expected contingencies may shape the sampling process. The results revealed that despite premature avoidance of certain choice options, self-determined information sampling can foster the proper use of the sampled information for contingency judgments. The likelihood of genuine contingency inference increasing if individual factors like the anticipated contingency, cognitive resources, or (reward-)structures of the environment facilitate sampling and integration of representative co-occurrence information.
Decision making under risk and uncertainty
Decision making under risk and uncertainty
Session chaired by Thorsten Pachur
10.30am
Good time, bad time: Do people invest processing effort adaptively in decision making with opportunity costs?
Kevin Tiede, Veronika Zilker, & Thorsten Pachur
When making choices, deliberation time is often costly but also promises to improve decision making. This, however, only applies if the available options differ in attractiveness, as additional processing time might not be worthwhile if the options yield similar value. We investigate to what extent people respond to the costs and benefits of time adaptively under varying value differences between options in decisions under risk. To study this trade-off between costs and benefits of time, we used the drift diffusion model (DDM) as a computational framework. Based on simulations, we first determined how people should adapt their boundary separation in order to maximize payoffs depending on the differences in options' value and the magnitude of opportunity costs (i.e., loss in payoff per time unit). Subsequently, we conducted three empirical experiments (total \(N = 453\)) to study if people adjust their processing adaptively, in line with the simulations. Across all experiments, participants' estimated boundary separations were mostly sensitive to the presence (vs. absence) of opportunity costs, but less so to different levels of costs and varying differences in option value. As a result, people did not adapt their boundaries to the value differences in an optimal fashion, with boundaries being narrower than optimal when value differences were relatively large and wider than optimal when value differences were relatively small. Although guessing (i.e., very narrow boundaries) would have been most beneficial in some conditions, participants did not adopt a guessing strategy in any condition. In conclusion, people seem to have limited abilities to adjust their information processing adaptively to the magnitude of the opportunity costs and value differences.
It is commonly assumed that people employ a range of different strategies for making decisions, including both simple heuristics and more complex strategies. Because these strategies differ in terms of which and how much information they rely on, their use might also be reflected in patterns in predecisional information search. But what are these strategies? And how variable is strategy use? Here we present a bottom-up approach to address these questions based on analyzing predecisional information search as a manifestation of the latent decision strategies. Using data from a process-tracing study with Mouselab, we used cluster analysis to identify patterns of information search in risky choice. The analysis yielded two clusters: One cluster ("balanced cluster") was characterized by a balanced inspection of both outcome and probability information with most transitions occurring between outcomes and probabilities within the same option. The other cluster ("outcome-priority cluster") showed a focus on inspecting outcome information and on transitions between different outcomes. The two clusters were associated with different choice behaviors; in addition, the choices associated with the balanced cluster were more consistent with the choice predictions of more complex strategies (e.g., least-likely), whereas the choices associated with the outcome-priority cluster were more consistent with the choice predictions of simpler strategies (e.g., maximax). Finally, variability in search seemed to be more strongly driven by individual differences between decision makers rather than by features of the choice problems. Our approach provides a bottom-up approach for investigating cognitive processes in risky choice and constraints current theorizing on the underlying strategies.
The development of formal models of decision making under risk has been shaped by how people decide between options with monetary outcomes. Whereas these models—with cumulative prospect theory (CPT) as its most prominent example—are successful in predicting such monetary choices, they provide a rather poor account of people's choices between risky options with nonmonetary outcomes (e.g., outcomes of medical treatments). Focusing on CPT, here we test psychologically motivated modifications of three basic assumptions underlying traditional models of risky choice: a) using affective ratings of the outcomes (rather than monetary equivalents) to quantify the outcomes; b) assuming context-dependent valuation of outcomes; c) assuming affect-dependent probability weighting. We submitted variants of CPT implementing these modifications to a rigorous model comparison. The results show a clearly superior performance of these variants when modeling choices with nonmonetary outcomes, but not when modeling choices with monetary outcomes. The best-performing model indicates that in choices with nonmonetary outcomes, the affective value of the worst outcome in a choice problem determines the processing of the entire problem. We discuss the implications of our findings toward a better understanding of the psychological underpinnings of nonmonetary risky choice, and how the processes differ between choices with monetary and nonmonetary consequences. We show that the conclusions also hold when implementing the modified assumptions in other modeling frameworks (e.g., decision field theory). Our results underline that some key assumptions of prominent models of risky choice do not generalize to decisions with nonmonetary consequences.
11.15am
Disentangling the impact of attribute and option attention on risky choice
Thorsten Pachur & Veronika Zilker
Probability weighting—the notion that risky outcomes are weighted as a function of their probabilities—is one of the most powerful theoretical ideas in descriptive models of risky choice and forms a central component of cumulative prospect theory (CPT; Tversky & Kahneman 1992). Recent empirical analyses suggest that probability weighting may be rooted in attentional processes: one analysis observed that differences in the shape of CPT's probability weighting function are associated with differences in how attention is allocated across attributes (i.e., probabilities vs. outcomes; attribute attention); another analysis observed a link between probability weighting and differences in how attention is allocated across options (option attention). These findings are perplexing, as they suggest that probability weighting functions with one and the same shape could be driven by very different attentional policies. Here we disentangle the links of probability weighting with attribute attention and option attention and investigate their relationship: Do they result from the two facets of attention allocation being empirically confounded; or do they represent distinct phenomena? Reanalyzing data from a process-tracing study with Mouselab, we find that biases in attribute and option attention are only weakly related to each other. More importantly, attribute and option attention have distinct and independent effects on probability weighting. The analyses contribute to a better understanding how of information attention shapes decisions under risk; they also underscore the limits of the explanatory value of CPT's probability weighting function. We discuss the implications of our findings for the development of cognitive process models of risky choice.
11.30am
Chimpanzee and human risk preferences show key similarities
Lou M. Haux, Jan M. Engelmann, Ruben Arslan, Ralph Hertwig, & Esther Herrmann
Risk and uncertainty are central to all adaptive decisions human and nonhuman animals make ‐ including when deciding where to forage, with whom to mate and how to deal with dangerous situations. Risk preference determines how agents decide in face of these hazards and is a crucial determinant of life outcomes such as health, wealth, and wellbeing. Yet, the biological underpinnings of this important building block of behavior remain unclear. Here, we investigated the extent to which chimpanzees' (Pan troglodytes) exhibit risk preferences that are in important respects isomorphic to human risk preferences. Chimpanzees are an ideal reference: they are one of humans' two closest living relatives, live in large, mixed sex social groups and have similar developmental stages with a long infant dependency, a period of adolescence and a long life expectancy of up to 50 years. Using a multimethod approach we studied chimpanzee (\(N = 86\)) risk preference across domains and measurements. Our results show that chimpanzees' willingness to take risks shares structural similarities with that of humans across four key dimensions: chimpanzee risk preference manifests as a trait-like preference that is consistent across domains and measurements, chimpanzees are ambiguity averse, males more risk prone than females and that the appetite for risk taking follows an inverted U-shaped relation to age. Our findings suggest that key dimensions of risk preference emerge independently of the influence of human cultural evolution and thus may have deeper phylogenetic roots than previously suspected.
Computational techniques that extract semantic representations of words from text corpora have shown promise in predicting risk perception. We investigate how an alternative route to semantic representations based on free associations compares to the text-based and the traditional psychometric approaches to risk perception. To this end, we recruited published data encompassing 307 risk sources and compared the out-of-sample predictive accuracy using a machine learning approach. The text-based approach employs fastText: a pre-trained model composed of 2M word vectors. The psychometric approach takes human judgements on nine dimensions of risk as input. Finally, the association-based representation was derived from over 3.6M free associations from the Small World of Words study. Firstly, we determined the pointwise mutual information between cues and responses. We then employed singular value decomposition to extract 300 dimensions that were comparable to those in fastText. We find that despite the association-based approach being trained on a dataset that is orders of magnitude smaller than fastText, out-of-sample accuracy for both models is equivalent, and comparable to traditional psychometric approaches. Furthermore, whilst dimensionality-reduced versions of the text-based approach see considerable decreases in out-of-sample performance, equivalent reductions of the association-based approach see no such decrease. Finally, notable gains in accuracy were achieved by combining the psychometric model with each semantic model, and these gains were greater when combining with the association-based representations than with the text-based. This suggests the risk-relevant information in free associations provides greater purchase on the psychometric approach than does text.
Judgment and decision making
Judgment and decision making
Session chaired by Agnes Rosner
4.00pm
Exemplar retrieval in preferential judgments
Agnes Rosner, Fabienne Roth, & Bettina von Helversen
Research on inferential judgments from multiple cues suggests that judgments are influenced by the retrieval of past instances (exemplars) stored in memory. Yet, on the process level little direct evidence exists that would allow a similar conclusion for preferential judgments, where there is no objective criterion to which a judgment can be compared. This study aimed to test if exemplar retrieval also plays a role in preferential judgments. In Experiment 1, half of the participants judged how much they would like smoothies consisting of different ingredients (preference condition) and the other half judged how much another person would like the presented smoothies (inference condition). In Experiment 2, all participants engaged in preferential judgments, but with or without instructions to respond as consistently as possible. To get direct evidence of exemplar retrieval, we recorded eye movements. Eye movements can be used to trace information search in memory, because when retrieving information, people look at spatial locations that have been associated with retrieval-relevant information but that are empty during judgment of new objects ("looking-at-nothing" behavior). The results show that people looked at exemplar locations in both inferential and preferential judgments, and both with and without instructions to respond as consistently as possible. Thus, the results suggest that people may rely on previously encountered exemplars also in preferential judgments and highlight the usefulness of studying eye movements "to nothing" to better understand the role of memory in judgment.
4.15pm
Preferences for intuition and analysis: Who, when and why?
Filipe Loureiro, Teresa Garcia-Marques, & Duane T. Wegener
People differ in their preferences to make decisions in a more intuitive or analytic manner. Despite evidence of reliable individual trait-like differences in intuitive and analytic styles, research has recently provided evidence for different ad hoc preferences for intuitive and analytic decision-making as a function of context characteristics. One of such characteristics is choice complexity. Evidence shows that choices that are perceived as complex elicit a greater preference for choosing analytically as opposed to intuitively, and choices that are perceived as simple elicit a greater preference for choosing intuitively as opposed to analytically (Inbar et al., 2010). Here, we test the direct and interactive effects of choice complexity and decision-makers' cognitive styles on explicit preferences for intuition and analysis (Study 1). Furthermore, we examine the role of perceived validity of intuition and analysis as mechanisms underlying such preferences (Study 2). Overall, results from both studies evidenced consistent explicit preferences for intuition in simple choices (purchase decisions of simple products) and analysis in complex choices (purchase decisions of complex products). Additionally, preferences for intuition (/analysis) were higher among more intuitive (/analytic) participants. Results from Study 2 provided evidence that individual differences in perceived validity of intuition and analysis mediated the effects of cognitive styles on preferences for intuition and analysis, being this effect clearer for complex choices. Taken together, these data suggest that both contextual features and individuals' cognitive styles influence preferences for decision-making strategies, which are partially guided by the perceived validity of such strategies.
4.30pm
Process accountability's potential to improve judgement and decision quality seems to be less robust than previously thought. An attempt to replicate the effect of (process) accountability on loss aversion in 3 studies
Johannes Rollwage, Stefan Schulz-Hardt, & Jan Alexander Häusser
Previous research has shown that accountability reduces loss aversion (Vieider, 2009), but no attempt has been made yet to differentiate this effect into process and outcome accountability ‐ a distinction that has been shown to be vital for the effects of accountability on judgements and decisions. We aimed to close this gap and extend the original findings by disentangling the two types of accountability. However, we were unable to do so, as over a series of three experiments we could not replicate the original effect of accountability on loss aversion. In Studies 1 (\(N = 127\)) and 2 (\(N = 90\)) participants were randomly assigned to one of three conditions: no accountability, process accountability or outcome accountability. Otherwise we employed the same task and experimental procedure as Vieider but were unable to replicate the effect of accountability on loss aversion as a Kruskall-Wallis test did not show any significant effects of experimental condition (Study 1: \(\chi^2 = 0.19\), \(p = .91\); Study 2: \(\chi^2 = 2.26\), \(p = .32\)). After our first two studies failed to replicate Vieiders results, our focus shifted towards a full-on replication attempt. Thus, for our third study (\(N = 183\)) we again used the same task and procedure as Vieider but this time removed the outcome accountability condition and applied Vieiders original accountability manipulation as closely as possible. However, we still were unable to replicate the original finding, as a Wilcoxon rank-sum test did not show significant differences in loss aversion between accountable and non-accountable participants (\(W\) = 4,514, \(p = .32\)). While our findings do not fully rule out that (process) accountability affects loss aversion, they strongly indicate that this effect, if it exists, is less robust than previously thought.
4.45pm
Unrealistic optimism is not quite what it seems when people are actually pessimistic about everyone else
5.00pm
Can conditionals explain explanations? A modus ponens model of "b because a"
Simone Sebben & Johannes Ullrich
We suggest a normative model for the evaluation of explanations B because A based on probabilistic conditional reasoning and compare it with empirical data. According to the modus ponens model of explanations, the probability of B because A should equal the joint probability of the conditional if A then B and the explanans A. We argue that B because A expresses the conjunction of A and B as well as positive relevance of A for B. In Study 1, participants (\(N = 80\)) judged the subjective probabilities of 20 sets of statements with a focus on belief-based reasoning under uncertainty. In Study 2, participants (\(N = 376\)) were randomly assigned to one of six conditions for which we varied the inferential relevance of A for B to assess boundary conditions of our model. We also report results on the Equation, a fundamental model in research on probabilistic reasoning concerning the evaluation of conditionals. In both studies, results indicate that participants' belief in statements B because A followed model predictions systematically. However, a sizeable proportion of sets of beliefs contained at least one incoherence, indicating deviations from the norms of rationality suggested by our model. In addition, results of Study 2 lend support to the idea that inferential relevance may be relevant for the evaluation of both conditionals and explanations.
5.15pm
The effect of diagnostic advice on clinical decision-making
Susanne Gaube, Harini Suresh, Eva Lermer, Marzyeh Ghassemi, Matthias Hudecek, Julia Dorn, Sabrina Rauh, Timo Koch, & Errol Colak
Artificial intelligence (AI) enabled clinical decision support systems are becoming more widespread in several medical fields, including radiology. However, comparatively little research has been conducted to evaluate the effect of such systems on physicians' judgment and clinical decision-making. In two pre-registered experiments, 256 and 223 physicians with varying task expertise received eight patient cases with chest x-rays and diagnostic advice. The advice was generated by human experts but labeled as coming either from an AI system or a human radiologist. Additionally, in Experiment 1, the advice's accuracy (correct vs. incorrect), and in Experiment 2, the advice's level of explainability (annotation vs. no annotation) were manipulated as within-subjects variables. In both experiments, participants were asked to evaluate the advice's quality and make a final diagnosis for each patient case. In Experiment 1, task experts rated the quality of advice lower when it came from an AI system, whereas non-task experts did not differentiate between the two sources of advice. Furthermore, the diagnostic performance was significantly worse when participants received inaccurate advice, regardless of the source of advice and task expertise. In Experiment 2, for which the data was collected two years later, neither the source nor the level of explainability of the advice affected the quality rating of the advice and diagnostic performance. However, even though all provided advice was correct in Experiment 2, the diagnostic performance was far from ideal. This work raises important considerations for the deployment of AI-generated advice in clinical settings.
Lavender Room
Facets and determinants of pro- and antisocial behavior
SymposiumFacets and determinants of pro- and antisocial behavior
Symposium organized by Angela Dorrough and Dorothee Mischkowski
Pro-social behavior is vital for the functioning of interpersonal relations, organizations, and societies at large. Correspondingly, the study of pro-social behavior and its counterpart ‐ anti-social behavior ‐ has gained considerable attention in all areas within psychology. In this symposium, we will bring together recent findings on different facets of prosocial behaviors and their determinants. Specifically, the symposium will present different facets of altruistic behavior (i.e., help-giving, moral courage, and peer punishment) as well as correlates of dispositional pro-sociality (i.e., benign and malicious envy). Furthermore, contextual facilitators (i.e., intergroup contact, shared group membership) and barriers (i.e., a bad reputation) of prosocial actions will be presented. Finally, we will discuss country-level differences in pro- and antisocial behaviors (i.e., social mindfulness and bribery behavior). Overall, the symposium offers an up-to-date summary of recent research on the composition of pro-social behavior and its determinants. Ultimately, it aims at encouraging the discussion about where the field stands and where it should head towards in the future.
8.30am
Social preferences and envy: Linking benign and malicious envy with social value orientation
Dorothee Mischkowski, Jens Lange, Jan Crusius, & Andreas Glöckner
Situations of disadvantageous inequality often elicit envy as a negative emotional response. To resolve it, an envious person can either strive to reach the superior status quo (benign envy) or can maliciously seek to restore equality so that the envied person loses her superiority (malicious envy). This project elaborates on the conceptual distinction between benign and malicious envy. Specifically, we investigate how benign and malicious envy relate to social preferences and develop an economic game to test whether dispositional benign and malicious envy predict envy-driven behavior. In two preregistered, incentivized online studies (\(N = 609\)), we (1) investigate the relation between Social Value Orientation (SVO) as a measure of social preferences and benign and malicious envy as individual difference measures and (2) validate the results in a newly developed economic game, mirroring benign and malicious envy as behavioral choice options. In both studies, we find a negative relation between SVO and malicious envy in that increased malicious envy goes in line with an individualistic preference. Unlike previous results, we find a similar relation between SVO and benign envy. When forcing individuals to decide either in favor of the benign or malicious choice option in the envy game, prosocials are more likely to choose the benign (vs. malicious) behavioral option. However, dispositional benign and malicious envy are not predictive of behavioral choice in the envy game. Our findings contribute to the recent debate on the nature of envy. In particular, the finding that benign (as well as malicious) envy negatively relate to SVO contradicts the concern that measures of benign envy might merely reflect "constructive" or socially desirable responses to inequality.
8.45am
Sustaining global public goods through intergroup interactions
Jörg Gross, Angelo Romano, Welmer Molenmaker, & Carsten de Dreu
Humans create and join groups in which they cooperate to tackle shared problems and increase joint welfare. Such group cooperation allows to establish club goods ‐ goods that benefit group members, while members of other groups can be excluded from their consumption and benefits. Many mechanisms, like peer punishment or reputation, have been shown to foster cooperation in small groups. However, being faced with global public goods problems, like climate change, the question arises how cooperation can emerge beyond confined groups, across group boundaries, and to the benefit of larger collectives. Using simulations, we show that frequent intergroup interactions play an important role in fostering global as opposed to group-based cooperation. We further show that decreased fragmentation of the population, i.e., a collective that is comprised of fewer sub-groups, increases the likelihood that true public goods rather than club goods are established. We further test core predictions of our model experimentally. Taken together, we identify why and when intergroup contact and reciprocal interactions across group boundaries are needed to establish cooperation that transcends group boundaries and allows to tackle global public goods problems.
9.00am
Social mindfulness and prosociality vary across the globe
Niels van Doesum, Ryan Murphy, Marcello Gallucci, & Paul Van Lange
Humans are social animals, but not everyone will be mindful of others to the same extent. Individual differences have been found, but would social mindfulness also be shaped by one's location in the world? Expecting cross-national differences to exist, we examined if and how social mindfulness differs across countries. At little to no material cost, social mindfulness typically entails small acts of attention or kindness. Even though fairly common, such low-cost cooperation has received little empirical attention. Measuring social mindfulness across 31 samples from industrialized countries and regions (\(n = 8,354\)), we found considerable variation. Among selected country-level variables, greater social mindfulness was most strongly associated with countries' better general performance on environmental protection. Together, our findings contribute to the literature on prosociality by targeting the kind of everyday cooperation that is more focused on communicating benevolence than on providing material benefits.
9.15am
Bribery within and across borders ‐ experimental evidence on conditional bribery from 21 nations
Angela Dorrough, Nils Köbis, Shaul Shalvi, Bernd Irlenbusch, & Andreas Glöckner
Curbing corrupt behavior like bribery presents a global grand challenge that increasingly spans national borders. Designing more effective anti-corruption policies requires understanding the behavioral underpinnings of bribery. To provide first behavioral insights into bribery within and across borders, we ran large incentivized online experiments using a bribery game with representative samples from 21 nations (total \(N = 6,472\)) across six continents. Results uncover differences between nations concerning the propensity to offer a bribe. Furthermore, indicative of what we label "conditional bribery" ‐ people offered bribes to interaction partners from nations with a reputation of being corrupt. Expectations about a nation's openness for bribery are shared across the participating nations but negatively correlate with reality.
9.30am
Constructing the Facets of Altruistic Behavior (FAB) scale
Sabine Windmann, Lucie Binder, & Martin Schultze
Behavior is effectively altruistic to the degree that it is costly for the actor while benefiting others. In a series of preregistered studies, we constructed a 15-item self-report scale assessing three different facets of altruistic behavioral traits: help-giving, moral courage, and peer punishment. Item selection was performed with the help of Ant Colony Optimization (ACO) procedures as implemented in the stuart package for R. Confirmatory factor analysis of the three-factor measurement model showed excellent fit, outperforming classical item selection procedures. The scale was structure-validated in a second sample using a multiple group model that showed full measurement and structural invariance. A pilot study shows correlations of the subscales with economic game decisions. We discuss the scale structure and potential applications.
9.45am
The more we know, the more we cooperate: The effect of shared group memberships on cooperation
Pinar Ugurlar, Onurcan Yılmaz, Ozan Isler, & Angela Dorrough
While the human psychological tendency to adopt an "us versus them" perspective has been fueling cooperating more readily with the in-group as compared to out-group members, the natural heterogeneity of social identities allows untapped opportunities for cooperation. Because people have multiple social identities, these identities tend to match across categories even though they may diverge in others. In the current line of research, we investigated how people cooperate across multiple social categories. In three preregistered experiments (total \(N = 1,092\)) participants with varying in- and out-group identity categories were paired in incentivized one-shot Prisoner's Dilemma (PD) game. The relationship between two participants in a pair was either in-group (two shared group memberships), mixed-group (one shared and one nonshared group memberships), or out-group (two nonshared group memberships) membership. Results consistently showed that cooperation increases with the number of shared group membership both in experimentally created groups (Experiment 1) and naturally existing groups (Experiments 2 & 3), and the effect of the shared group membership on cooperation was mediated by the identification with the interaction partner but not mediated by the expected cooperation from the interaction partner. The consistency among experiments suggest that the results are not driven by specific in-group/out-group considerations, but rather that the observed disparities in perceptions that are evoked by shared group membership.
Attitudes and stereotypes
Attitudes and stereotypes
Session chaired by Tal Moran
10.30am
The prospect of long-term implicit attitude change: Tests of spontaneous recovery and reinstatement
Benedek Kurdi, Thomas Mann, Jordan Axt, & Melissa Ferguson
Implicit (automatic) evaluations can exhibit momentarily malleability, including in the negative-to-positive direction. However, demonstrations of long-term change are exceedingly rare, especially in the context of real-world targets. In this project (\(N\) = 2,150), we rely on ideas from fear leaning to ask why even interventions that produce short-term malleability immediately tend to be ineffective in creating durable change in implicit attitudes. In Experiment 1, we identified two interventions that were particularly effective in overturning negative implicit evaluations formed about a novel social target in the short term: reinterpretation and negation + replacement. In Experiment 2, we confirmed the effectiveness of these two interventions in bringing about immediate reversals of experimentally induced negative implicit evaluations. In Experiment 3, the shifts in implicit evaluation achieved via reinterpretation and negation + replacement did not show any decrement over 2 days, thus providing evidence against spontaneous recovery. However, when the target was repeatedly paired with mildly unpleasant images 2 days after reinterpretation or negation + replacement (reinstatement), implicit evaluations became markedly negative (Experiment 4) and considerably more negative than implicit evaluations of a control target that had never been subject to implicit negativity (Experiment 5). Taken together, these data suggest that negative-to-positive changes in implicit evaluation may not be vulnerable to the mere passage of time. However, even interventions that seem to fully and durably overturn implicit negativity can leave initially encoded negative information intact and allow for such information to be expressed later, including in response to subtle reminders.
10.45am
Attitude similarity and interpersonal liking: A dominance of positive over negative attitudes
Tabea Zorn, André Mata, & Hans Alves
Sharing attitudes leads to liking. Dating apps and other social media platforms regularly exploit this phenomenon by matching dating partners based on shared attitudes or interests. While this similarity effect is well-established, past research rarely addressed whether it matters if the attitudes are positive or negative; that is, whether positive and negative attitudes differ in their potential to elicit liking. Hence, it is unclear whether people prefer others who share their likes or others who share their dislikes. While the sparse research that has been done on this topic suggests that shared dislikes may evoke stronger liking (Bosson et al, 2006; Weaver & Bosson, 2011), we predicted that shared likes have a stronger potential to elicit liking than shared dislikes in the initial stages of impression formation. We based our assumption on principles of impression formation such as people's general preference for others who express likes over others who express dislikes (Burnstein, 1967; Folkes & Sears, 1977). In four studies (\(N = 402\)), we showed that likes have a stronger potential to elicit liking than dislikes. That is, participants found others who shared their likes more likable than others who shared their dislikes (Study 1). Also, participants found others who did not share their likes least likable, while not sharing dislikes was not as detrimental to liking (Study 2). We argue that three aspects contribute to this finding. First, people generally prefer likers to dislikers (Study 3). Second and third, likes are stronger and more self-revealing than dislikes (Studies 2 & 4). Our work offers novel insights into the similarity effect and has implications for dating and friendship initiation.
11.00am
Increasing the predictive power of attitude implicit association tests by introducing the difficulty concept from classical test theory
Merlin Urban, Tobias Koch, & Klaus Rothermund
We propose a test theoretical approach to tackle the issue of low predictive power of attitude implicit association tests (IATs) by introducing the difficulty concept from classical test theory (CTT). Following CTT, we argue that attitude IATs of medium difficulty (defined as mean IAT scores of zero) have more predictive power than attitude IATs of extreme difficulties (defined as mean IAT scores deviating strongly from zero). This relation between IAT difficulty and predictive power is assumed to be mediated by the variance in IAT scores, with moderate difficulty resulting in more variability. To test our hypotheses, we compared IATs of different difficulties with regard to their ability to predict direct attitude measures, drawing on the Attitudes, Identities, and Individual Differences Study (AIID), a large experimental online study that ran on project implicit. In total 95 different attitude IATs and 127,259 complete cases went into the analyses. In order to account for the hierarchical data structure, we translated our hypotheses into multilevel models and tested them using multilevel structural equation modelling. In line with our hypotheses, IAT difficulty proved to be a powerful moderator of the IATs predictive validity, and the effect of difficulty was mediated by variance. The results suggests that striving for strong and robust IAT effects, i.e. IATs of extreme difficulties, which seems to be common practice in IAT research, actually reduces the predictive power of IATs. We recommend that researchers interested in using the IAT to predict outcomes develop IATs of medium difficulty in order to maximize variability in IAT scores. We also make suggestions how to manipulate IAT difficulty.
11.15am
Maintenance of stereotypical expectations through contextualization of stereotype-incongruent behavior
Kiara Roth & Sarah Teige-Mocigemba
To date, several psychological mechanisms contributing to the maintenance of stereotypical beliefs have been identified. One of these mechanisms constitutes the contextualized processing of information which contradicts stereotypical beliefs, that is, the incorporation of contextual cues into the representation of an object, person, or group. Using a modified predictive learning paradigm, the present study (\(N = 369\)) investigated the contextualization of stereotype-incongruent behaviors shown by either a member of a German student fraternity or a member of a politically left group. In the learning phase, participants formed associations between the target person's behaviors (congruent or incongruent) and the people accompanying him, representing the social context. To this end, we manipulated the predictive value of the context persons such that the target's (in)congruent behaviors were perfectly associated with the presence of one context person, respectively, while another context person was present independent of whether the target behaved (in)congruently. In the subsequent test phase, we varied the social context and assessed the proportion of predicted stereotype-congruent or incongruent behaviors depending on who the target was with. Results showed that participants formed context-dependent expectations of participants' behaviors, and that behavior predictions were influenced by participants' political identity. We discuss the role of motivated contextualized information processing for the maintenance versus change of stereotypical expectations.
11.30am
The impact of counter-stereotypical exposure on automatic stereotyping depends on validity information
Tal Moran, Femke Cathelyn, Pieter Van Dessel, & Jan De Houwer
Whereas automatic stereotyping is typically attributed to the activation of associations, we recently proposed that it depends on the activation of propositional representations. We tested the crucial prediction of the propositional perspective that validity information can moderate automatic stereotyping. More specifically, we examined whether the effectiveness of exposure to counter-stereotypical individuals, a well-studied intervention to reduce automatic stereotyping, is moderated by validity information. A pilot study established an intervention that influenced automatic stereotyping through exposure to descriptions of women in leadership and men in supporter positions (i.e., counter-stereotypical exposure) or vice versa. A pre-registered experiment (\(N = 377\)) found that the effect of this intervention was moderated by information about the accuracy of these descriptions (i.e., validity information). We discuss how these findings can inform research on automatic stereotyping and interventions aimed at reducing such stereotyping.
Trust in science—across all political camps—is crucial for modern societies. However, previous work suggests that a conservative ideology may, in general, be detrimental to trusting science. Here, we challenge the idea that any specific ideology is necessarily connected to (dis)trusting science. Instead, we argue that, across the political spectrum, trust and distrust in science result from specific stereotypes about scientists. Building on the ABC model of stereotypes about groups, we hypothesize that people hold spontaneous stereotypes about scientists' political beliefs, which are compared to one's own political beliefs. This comparison is, then, used to decide whether to trust or distrust scientists. Put simply, we assume that people trust scientists when they believe that scientists have a similar political ideology to themselves. A first experiment (\(N = 200\) US Americans), in which we manipulated scientists' perceived ideology (conservative versus liberal) fully supports this hypothesis. When scientists were perceived to be liberal, liberal participants reported greater trust in science, similar to previous findings. However, when scientists were perceived to be conservative, this effect fully reversed and now conservative participants showed greater trust in science. These findings suggest that stereotypic beliefs about scientists' ideology may strongly matter for trust in science. As the next step, we are now running a series of four additional studies to explore the potential boundaries and consequences of our findings (e.g., for the perception of specific scientific disciplines).
Pine Hall
Experimental aesthetics I: Following Fechner’s conceptions
SymposiumExperimental aesthetics I: Following Fechner’s conceptions
Symposium organized by Thomas Jacobsen
The year 1876 marks the beginning of second-oldest branch of Experimental Psychology, namely Experimental Aesthetics. In his major work Vorschule der Ästhetik, Gustav Theodor Fechner suggested the empirical, experimental study of aesthetics "from below", applying empirical knowledge. To date, the psychology of aesthetics has been the subject of increasing interest. The present symposia comprise contributions investigating a variety of aesthetic domains including, inter alia, music, materials, and visual artworks. Furthermore, the researchers address questions of the influence of several stimuli's and individual's characteristics, for example, complexity and memory resources, respectively.
8.30am
Cross-cultural empirical aesthetics of dance
Julia F. Christensen, Shahrzad Khorsandi, & Winfried Menninghaus
Cross-cultural empirical research in all domains of study of human behavior, mind and brain has been programmatically called for in a seminal article by Henrich et al. (2010) [Henrich, J., et al. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2-3): 61-83; discussion 83-135.]. Research in empirical aesthetics has recently started to embrace this call in a systematic manner. In this talk, we present the Persian Dance Movement Library (PDML), along with extensive cross-cultural online validation data from perceptual rating tasks with human observers from two cultures (Iranian, English; \(N = 322\)). The PDML contains 20 custom-choreographed dance sequences of six-seconds length (8 counts), danced five times each, with a different emotional intention by the dancer at each repetition; joy, anger, fear, sadness, and a control condition, neutral expressivity (stimuli: \(N = 100\)). In four separate cross-cultural online rating experiments, Emotion Recognition (ER) rates, along with Expressivity, Beauty and Liking ratings were obtained. While ER rates were relatively low, ratings of Expressivity, Beauty and Liking differed significantly between the emotion categories, suggesting that, perceptually, participants differentiated the emotional intentions of the dancer, even if labelling these according to classical emotion categories was difficult. To investigate the effect of cultural familiarity, a Culture Questionnaire was developed, and its factors as covariates showed that enculturation modulated all perceptual judgments. Results are discussed in the context of research from the domains of general emotion psychology and of cross-cultural empirical aesthetics.
8.45am
Concerts as an influencing factor of music experiences: Evidence from live and streamed concerts
Julia Merrill, Anna Czepiel, & Melanie Wald-Fuhrmann
While experimental aesthetics has long worked primarily with artificial stimuli in laboratory contexts, researchers have begun to acknowledge the need to study responses to actual artworks in realistic perception contexts. In our group, we take up this challenge by examining music listening in live concerts. We seek to determine which aspects of a concert influence which aspects of the musical experience. In a first exploratory study, we measured aesthetic experience in combination with psychophysiology from 98 participants during three live concerts with an identical programme of string quintets. Different analysis methods were applied to compare the acoustics and the self-reports with the physiological responses during the entire concert. With regard to acoustics, inter-subject correlation was used to identify reliable responses to musical events. It was found that highly correlated responses depicted typical signatures of physiological arousal, but also occurred during important structural moments in the music, namely at transitional passages, boundaries, and phrase repetitions. In a second study, we measured self-reports of aesthetic experience in response to four types of streams of the same musical programme. Stream type had an effect on aspects of general concert experience and on piece-evoked aesthetic emotions. However, these effects were often mediated by stream-type preference, which we had identified via a latent profile analysis. In particular, streams that either included the opportunity for social interaction of audience members or additional information on the programme and the contemporary piece were able to level out negative effects of a group of less interested audience members.
9.00am
A playlist from the past: Musical associations and nostalgic feelings
Marius H. Raab, Alicia Biener, Charlotte Habbig, Karim Weth, & Claus-Christian Carbonpresenting
Music is able to elicit strong emotions and can be used to induce feelings of nostalgia, and to prime biographic memories that are formative for one's self concept. This might have beneficial effects for finding and pursuing long-term goals in life (Sedikides et al., 2018). In two separate studies, we examined musical phenomena from popular culture in relation to nostalgic feelings. Study 1 (\(n = 40\)) used the title songs from five very popular children's TV series from the 90s. People between 19 and 67 years listened to the songs and were, after each song, asked for ratings, verbal assessments and associations related to those songs. In Study 2, 46 people who were deliberately drawn from a narrow age range (23 to 27 years) were asked to compile their own playlist ‐ the contemporary equivalent of a mixtape ‐ of five songs that "remind you of your own youth". They answered ratings and open questions regarding their associations after listening to each song. Also, their facial expressions were recorded on video and assessed with the FaceReader software. We demonstrated in both studies that biography-related music has a potential for evoking strong and positive emotions. Both studies also showed that, depending on person and context, mixed and "bittersweet" emotions are integral part of the experience. By associating past emotions and experiences with distinct tunes, they get accessible for a present context—in the words of Gustav Theodor Fechner, amalgamating with the "sensual impression, thus enriching it, illustrating it". We also found that a multi-methodological approach of quantitative and qualitative assessments, supported by an automated evaluation of facial expressions, is expedient to grasp the manifold experience of connecting to one's earlier self via music.
9.15am
A computational model predicts individual aesthetic judgments
Aenne A. Brielmann, Max Berentelg, & Peter Dayan
Where do you want to live? With whom? Do you like that image or swipe it away? Numerous decisions, big and small, partly depend on options' sensory appeal. Yet, we have a poor understanding of how sensory experiences gain value and how these values influence our decisions. We propose the theory that observers maintain and adapt the states of their cognitive-sensory system in order to process stimuli effectively now and in the future. Two interlinked components generate an object's sensory value: 1) processing fluency ‐ the likelihood of a stimulus given an observer's state; 2) the change in fluency with which likely future stimuli will be processed ‐ the change in the average likelihood of expected future stimuli. We test a simple realization of this theory that represents current and expected stimulus likelihoods as n-dimensional Gaussians. To control the position of naturalistic stimuli in this n-dimensional space, we created 55 morphed images from seven original pictures. Our participants (\(N = 59\)) first rated how much they liked each image. Then, they spent 15 minutes viewing the images, one at a time in random order, repeating after all 55 had been shown, while being given the option to move on to the next image whenever they wanted to. Finally, they rated all images again. We fit our model and alternative models to the first ratings. We then predicted the second ratings based on model fits and viewing times from the free-viewing part of the experiment. Our model explains, on average, 50% of the variance in the second ratings, and it outperforms regression models with equivalent predictor variables in most cases (40/59 participants). In brief, we show that our computational model of sensory value can predict individual aesthetic judgments and their change over time.
9.30am
Investigating the genuineness effect: Looking at real works of art vs. reproductions
Eva Specker & Helmut Leder
As our world becomes more digital, our interactions with art increasingly occur online through reproductions. In the psychological literature, the difference in aesthetic experience between a physical work of art and its (digital) reproduction is referred to as the "genuineness effect". This effect will be the topic of discussion of this talk. Specifically, we will first review a recent meta-analysis (Specker et al., 2021) that essentially found no evidence for this effect. We then discuss why a genuineness effect has not been found so far, and then discuss empirical work that tries to tease apart these reasons. Specker and Leder (under review) focus on the facsimele accommodation hypothesis proposed by Locher et al. (1999, 2001) and Specker, Arato, and Leder (under review) focus on the well-known anchoring effect as a potential confound. As our world becomes more digital, and especially now, in the wake of the Covid-19 pandemic, people often have not been able to engage with physical artworks and all kinds of cultural institutions have been forced to close their doors, and are only accessible virtually, the question of what real artworks contribute to our aesthetic experiences becomes more and more pressing. Therefore, a better understanding of the genuineness effect—or lack thereof—could not only have a far-reaching impact on the role of cultural institutions and art in our society, but also our scientific understanding of how we interact and engage with art as well as our empirical approach to studying these topics. We hope in this talk to take a first step towards furthering our understanding of this topic.
9.45am
Disentangling individuals' aesthetic preferences for order and complexity: A parametric, multidimensional approach
Eline Van Geert, Daniel Hofmann, Astrid Warny, & Johan Wagemans
Which visual displays do we aesthetically appreciate, and why? Order and complexity have often been suggested as important factors related to appreciation, but the exact type and direction of their relation to appreciation is still under debate. To disentangle individuals' aesthetic preferences for order and complexity, it is important (a) to manipulate order and complexity independently and in a parametrically controlled fashion; (b) to study the relation of both factors to appreciation in combination rather than separately; and (c) to acknowledge the multidimensional nature of order and complexity. In a first study, we used the recently developed OCTA toolbox to investigate whether individuals' preferred complexity levels for color, shape, and size depend on the level of order present in the display. In a 2AFC task, participants (\(N = 415\)) indicated their preferences for a series of image pairs, in which complexity levels were varied within each pair, and order levels between pairs. Also, individual differences in preferences for color, shape, and size complexity were explored. Results indicated that (a) not all complexity types are appreciated similarly, (b) not all individuals appreciate complexity equally, and (c) complexity is more often preferred in highly ordered stimuli than in stimuli with a lower order level. A second study examined ratings of order, complexity, pleasantness, and interest for a parametrically controlled set of 1,611 stimuli varying systematically in multiple order and complexity dimensions. Results confirmed the differential relation of different types of appreciation to order and complexity, and provide insight in the extent to which some objective order and complexity manipulations influence perceived order and complexity.
Experimental aesthetics II: Following Fechner’s conceptions
SymposiumExperimental aesthetics II: Following Fechner’s conceptions
Symposium organized by Thomas Jacobsen
The year 1876 marks the beginning of second-oldest branch of Experimental Psychology, namely Experimental Aesthetics. In his major work Vorschule der Ästhetik, Gustav Theodor Fechner suggested the empirical, experimental study of aesthetics "from below", applying empirical knowledge. To date, the psychology of aesthetics has been the subject of increasing interest. The present symposia comprise contributions investigating a variety of aesthetic domains including, inter alia, music, materials, and visual artworks. Furthermore, the researchers address questions of the influence of several stimulus and individual's characteristics, for example, complexity and memory resources, respectively.
10.30am
Colour and contour features convey emotions in abstract visual artworks
Claudia Damiano, Pinaki Gayen, Archi Banerjee, Gobinda Banik, Priyadarshi Patnaik, Dirk B. Walther, & Johan Wagemans
Visual abstract art is often used to convey feelings and emotions through the manipulation of colour and form. Here, we investigated how colours and lines are used to express emotions through art. A secondary question was whether artists and non-artists express emotions through art in similar ways. In our study, 46 artists and 45 non-artists each created abstract colour drawings and line drawings depicting six emotions (i.e., anger, disgust, fear, joy, sadness, and wonder). To determine whether emotions are depicted in common ways across individuals, we computationally predicted the emotion of a given drawing by comparing it (i.e., pixel-wise correlation) to a set of reference drawings created by averaging across all other participants' drawings within each emotion category. Prediction accuracy was above-chance in all cases, but was higher for colour drawings (accuracy = 43.8%) than line drawings (23.6%, \(p < .001\)), and for non-artists' drawings (36.9%) than artists' drawings (30.6%, \(p < 0.05\)). Additionally, in a separate behavioural experiment, we found that people (\(N = 242\)) could also accurately infer emotions, showing the same pattern of results as our computational predictions. Further analyses of the drawings revealed that there are systematic commonalities in colour and line usage when depicting an emotion, and differences across emotions (e.g., anger is generally red, and more densely drawn than other emotions). We also found that artists tend to use darker lines and fewer colours than non-artists to depict emotions. Taken together, these results imply that abstract colour and line drawings convey certain emotions based on their visual features, which are also used by humans in the production and appreciation of abstract artworks.
In his book 'Analysis of Beauty' from 1753, the British artist William Hogarth introduced his 'Line of Beauty' (LoB) by depicting seven S-shaped lines and declaring #4, which has a certain curvature, as the most beautiful. Even today, the LoB has a persistently strong influence on many areas such as landscape art and design, calligraphy, furniture design, architecture, dance, etc. Therefore, it is all the more astonishing that Hogarth's claim has never been tested empirically. In our study, we made up for it and found that the LoB is indeed the most preferred among Hogarth's seven lines. However, since the LoB also differs from the other lines in several respects, such as thickness, length, etc., it is open whether its preference was due to its specific curvature alone. To investigate the effects of curvature and other properties on aesthetic preference, we conducted further experiments in which these properties were varied systematically. Our results show that preference is nonlinearly related to curvature. Moreover, S-shapes whose lower arc is slightly longer than the upper arc, are preferred. Together, these two properties explain more than 80% of the variance in preference judgments.
Finding beauty in everyday artefacts, evaluating them, and deriving feelings of pleasure from the whole process is a universal cognitive function in humans that has been proven adaptive during species evolution. Such pervasiveness and universality of beauty judgments suggest the existence of biological substrates for them that can be systematically investigated. One of the most common everyday artefacts is music; when listening to music, even without any conscious effort or specialized education, people judge whether a musical piece is beautiful or not, namely they issue an aesthetic judgment. Up to now only few studies have focused on elucidating the neural mechanisms that govern this process, and among them, neuroimaging experiments by our and other labs identified the medial orbitofrontal brain region as robustly associated to judging either music or paintings as beautiful. For music, neural communication with the supratemporal lobe is required. Our recent magnetoencephalography research further showed that the memorization of aesthetically valuable, tonal musical patterns by J. S. Bach strengthened large-scale connections between auditory areas, basal ganglia, cingulate and hippocampal regions. In turn, the memorization of atonal, non-aesthetically pleasing patterns was related to a more focal auditory-cortex activity with lower extrinsic connectivity. Moreover, we analyzed the dynamic functional connectivity related to listening to musical patterns typically judged as beautiful with that related to listening to patterns judged as ugly and found recurrent brain states switching between auditory, orbitofrontal, insular, and striatal networks. Hence, the wide, complex interplay between brain regions seems to be the brain mechanism underlying positive appraisal during a musical aesthetic experience.
The experience of being absorbed in music is a particular aesthetic experience that has been investigated with a variety of self-report scales. Absorption has been related to attentional capacity directed at processing an aesthetic stimulus. The goal of our research is to relate self-rated musical experiences to ocular measures, such as pupil dilation, blinking, and saccadic activity. Besides aesthetic absorption, we are interested in states and emotions like mind-wandering, groove, valence, arousal. We will report an overview of our studies from the last years. We recorded eye measures during music listening in varied settings, differing in instructions of fixation (fixation on central target or free viewing on an empty screen) and blink behavior (inhibited or spontaneous). On the coarse level of analyzing behavior on trial means, we found relations between feeling absorbed and mean saccadic activity, blink rate, and blink durations. Musical features also predicted eye parameters and the explained variance was independent of subjective states. On a more fine-grained level, taking the dynamics of the ocular measures into account, pupil dilation was related to musical structure and expectation violation. Further, various features of the pupil signal can predict different states. For example, phase coherence predicts absorption, while the number of pupil size changes predicts feeling in the groove. Most of our results have been replicated across three studies, though some show sensitivity to context settings.
11.30am
Self-relevance, but not goal-relevance, increases aesthetic appeal of synthetic artworks
Scherien Müller, Ayse Ilkay Isik, & Edward A. Vessel
Aesthetic experiences with artwork can be highly impactful, yet it is not well understood why some artworks touch us more than others. In previous work, we found that self-relevance, defined as whether an artwork relates to one's self-identity, interests and past experiences, was a strong predictor of aesthetic appeal. We hypothesized that the presence of personally relevant information in artworks allowed for deeper processing and hence greater aesthetic appeal. In this experiment, we sought to test whether a different type of relevance, goal-relevance (does the content of the artwork relate to current goals or tasks), would have a similar effect. We manipulated goal-relevance using an ownership task. During the training phase, participants learned the assignment of 20 different objects as either self-assigned ("Mine") or as other-assigned with a gender neutral equivalent ("Pat's"). Better performance and faster reaction times for identification of self-assigned items suggested that, on average, the task was successful at creating a tighter link to the self-assigned items. In a second phase, observers viewed a set of synthetic artworks that depicted the objects from the ownership task, intermixed with a set of real artworks. In a first block, the participants rated how aesthetically moving they found each artwork. In a second block, they rated the same set of artworks for self-relevance. Synthetic artworks containing objects from the self-assigned versus other-assigned conditions did not differ on either aesthetic ratings nor judgments of self-relevance. This suggests that unlike links to autobiographical memory, identity and interests, manipulation of goal-relevance via the ownership task was not sufficient to engage deeper processing and affect aesthetic appeal.
The present study aimed to examine whether the circadian rhythm modulates our aesthetic judgement and explore the potential factors involved in such judgments. Eighteen participants were asked to rate the relative aesthetic values of a pair of paintings, artworks from a famous Chinese painter, once every two hours from 8:00 am to 10:00 pm. The results showed that time of day has a tendency to modulate the aesthetic judgement, however with inter-individual differences. The relative aesthetic judgements decreased significantly in the afternoon compared to morning hours and rebounded in the evening, with the second relative judgement compared to the first of the pair pictures being judged most "beautiful" at 2:00pm but "least beautiful" at 6:00pm. Among the examined factors, color and shape played most important roles in aesthetic judgement. Our results indicate that the circadian rhythm modulates our aesthetic experience, and further highlights the necessity to consider the time of day when investigating perceptual and cognitive functions.
Experimental aesthetics III: Following Fechner’s conceptions
SymposiumExperimental aesthetics III: Following Fechner’s conceptions
Symposium organized by Thomas Jacobsen
The year 1876 marks the beginning of second-oldest branch of Experimental Psychology, namely Experimental Aesthetics. In his major work Vorschule der Ästhetik, Gustav Theodor Fechner suggested the empirical, experimental study of aesthetics "from below", applying empirical knowledge. To date, the psychology of aesthetics has been the subject of increasing interest. The present symposia comprise contributions investigating a variety of aesthetic domains including, inter alia, music, materials, and visual artworks. Furthermore, the researchers address questions of the influence of several stimulus and individual's characteristics, for example, complexity and memory resources, respectively.
4.00pm
Preference for symmetry: From Fechner to modern empirical aesthetics
Andreas Gartus & Helmut Leder
Almost 150 years ago, Gustav Theodor Fechner acknowledged the fact that symmetry is generally preferred over asymmetry. More specifically, he proposed that preference is not proportionally decreasing with deviation from symmetry. Instead, he assumed that as soon as the deviation is big enough to be recognized, it will cause a strong decrease of preference. This is exactly what we found in a first study: Abstract patterns showing small random deviations from symmetry (broken patterns) were profoundly less liked than symmetric patterns. However, researchers from Bertamini Lab have found a more linear relation between proportion of symmetry and preference, which could be due to a different way of constructing levels of symmetry in stimulus patterns. With these conflicting results, the influence of deviations from symmetry on liking was still unclear. Therefore, in another study using Fechner's method of production, we instructed participants to change abstract patterns randomly deviating from symmetry (broken patterns) to improve their aesthetic appeal without making them symmetric. A follow-up rating study showed that the improved patterns were almost as liked as symmetric ones. In addition, they were also rated much more interesting than symmetric patterns. Furthermore, there are also large individual differences. In a recent study, we compared art experts with laypersons and found no difference in liking or interestingness between symmetric and asymmetric patterns for art experts. Thus, we could only partly confirm Fechner's original statement about the negative effect of deviations from symmetry on preference. It seems to be correct for laypersons and random, clearly visible, focal deviations, but not necessarily for intentional, balanced deviations or art experts.
Aesthetic processing has a great impact on our everyday lives. It influences our choices regarding romantic partners, where we wish to live, how we dress, which objects we surround ourselves with, and the activities we pursue in our leisure time. Aesthetic considerations affect health, productivity, and learning; and participation in cultural activities is positively related to wellbeing. The desire to identify the principles underlying aesthetic processing has played a dominant role in the history of human thought. For more than two millennia, theoretical reflections on aesthetic perception have been an important domain of philosophy. By contrast, empirical approaches to aesthetics ‐ as first advocated in mid-nineteenth century psychology ‐ have not found any major, sustained representation at universities and are only now on the verge of becoming accepted as a mainstream field in experimental psychology and the cognitive neurosciences. But what actually is the agenda of empirical aesthetics? What are its main goals? What characterizes aesthetic experiences? Why is it important to study them, and how? This presentation will touch upon these issues. It will discuss how research conducted so far can be categorized as either being subject-oriented or stimulus-oriented. It will argue that for its further development, empirical aesthetics must integrate these approaches into a unified paradigm with a strong focus on temporal unfolding and on interactionist processes within both the perceiver and the stimulus. Finally, the need of conceptual clarity will be addressed, the scope of empirical aesthetics and its integration into a broader academic landscape.
4.30pm
Art forms in nature: Perceptual and aesthetic properties of Ernst Haeckel's drawings of new species
Johan Wagemans, Ines De Vlieghe, Christophe Bossens, Derya Soydaner, Claudia Damiano, & Eline Van Geert
Most research in empirical aesthetics belongs to one of two camps: (1) focus on the relation between beauty and simple quantitative properties in well-controlled stimuli, (2) including a broader range of stimulus and aesthetic variables using more ecological stimuli such as natural images and art works. Ernst Haeckel's drawings of "Kunstformen der Natur" (1899‐1904) are ideally suited as a bridge between these two traditions: They are plates consisting of a variable number of separately drawn art forms (e.g., plants, shells, animals), with different grey levels and colors for figures and backgrounds, with a good level of order (e.g., symmetry) and complexity (e.g., dense etchings) ‐ all beautifully rendered and composed. In a large online study, we presented high-resolution scans of the 100 prints and two sets of 100 isolated art forms extracted from 16 prints to different groups of participants (total \(N = 203\)), which they had to rate on 7-point Likert scales for beauty, pleasure, interest, order, and complexity. In addition to basic demographics, all participants indicated their familiarity with biology and Haeckel, interest in art, connectedness with nature, and inclination to "otherworldliness". Beauty, pleasure and interest were strongly positively inter-correlated, while order and complexity were correlated negatively. Of the three aesthetic scales, interest was most strongly influenced by order and complexity. Average aesthetic ratings were higher for the whole plates than for the separate items ("parts"), across all stimuli as well as for specific wholes and parts from the same plates. Additional analyses will investigate relations with objective measures of order and complexity, as well as possible group differences as a function of some individual characteristics.
4.45pm
Conceptual structure of materials aesthetics ‐ what about expertise?
Barbara E. Marschallek & Thomas Jacobsen
Following Fechner (1876), we drew on a bottom-up, exploratory "aesthetics from below" approach to elucidate non-experts' and experts' conceptual structure underlying the aesthetics of different materials (Werkstoffe). This was done for ceramics, glass, leather, metal, paper, plastic, stone, textiles and wood, as well as for the superordinate category materials. Adopting a technique first used by Jacobsen, Buchta, Köhler & Schröger (2004), we asked 2452 non-experts, as well as 401 experts (architects, designers, interior designers) to list adjectives one could use to describe the aesthetics of the materials with two minutes time on task. Descriptive, sensorial qualities constituted the core elements of the non-experts' conceptual structure of the aesthetics of materials, with an outstanding reference to haptic qualities, for example, smooth and cold. Results of an additional valence rating study ascribed the listed terms a rather neutral valence (Marschallek, Wagner & Jacobsen, 2021). Besides a large number of related adjectives, experts additionally listed aesthetically evaluative terms, for example, modern and elegant. As measured by the ratings in a second supplementary study, a slight majority of the listed terms had a neutral valence, followed by a large proportion of positive terms. The term beautiful, frequently the preeminent term in various domains of aesthetics, was by no means one of the most relevant terms in both studies. The present results show that the conceptual structure of the aesthetics of materials is diversified and rich, and, to some extent, influenced by expertise.
5.00pm
Reduced working memory resources affect aesthetic experiences
Rosalie Weigand & Thomas Jacobsen
Do we savour aesthetic experiences less when distracted by interrupted tasks, work rumination, or stress? Evidence suggests that the ability to concentrate on the aesthetic experience is crucial for initiating a processing mode of conscious aesthetic reception that results in more positive emotions. When working memory resources are otherwise occupied, people are less able to concentrate on aesthetic experiences. We conducted three investigations to examine how conditions that are known to deplete working memory resources affect the savouring of aesthetic experiences. In a controlled laboratory setting, participants rated beauty and savouring felt from encounters with visual stimuli after an interruption of a writing task. Aesthetic experience was hampered if participants were interrupted. In two field investigations, we demonstrated that work-related rumination and stress are inversely related to the savouring felt from opera, theatre, or cabaret pieces. These findings highlight the importance of concentrating on aesthetic experiences so that the perceiver can fully benefit from them.
5.15pm
Do internalized social norms and expertise influence how subpopulations appreciate tattoo aesthetics? The conceptual structure of tattoo aesthetics
Selina Maria Weiler & Thomas Jacobsenpresenting
Body modifications find more and more acceptance and popularity. Therefore, the question arose, how different subpopulations appreciate tattoos in terms of aesthetics. The goal of this study was to investigate the conceptual structure of tattoo aesthetics, with a particular focus on the effects of internalized social norms and expertise. Three groups (\(\leq 49\) years, \(\geq\) 50 years, and experts) of 497 participants were instructed in a timed free-listing task to write down adjectives that could define tattoo aesthetics. Applying several statistical analyses including a generalized Procrustes analysis, we highlighted the variance and complexity with which individuals verbally expressed their perceived aesthetic appeal of tattoos. Interestingly, for two of the three groups, the results did not reflect a coherent concept of beauty, nor did they present a clear bipolar dimension of beautiful/ugly. Nonetheless, the concept of beauty was found prevalent in tattoo aesthetics, as well as the identification of aesthetic and descriptive‐evaluative dimensions, with terms such as beautiful, ugly, multicolored, and interesting being the most notable adjectives, although not with the highest valence. Based on the intricacy of the phenomenon, the study discovered a high overall diversity of terms that elicit cognitive and affective components as well as uncovered results influenced by person‐situation interactions in a field-theoretic sense.