Mer de Glace View

When you want to post your talk title, abstract, and authors, email Rich Shiffrin who will send you the code to enable you to do so.

Authors, Titles, Abstracts, Presentations

IMPORTANT:
The ASIC speakers and attendees, whether world famous scientists or graduate students, expect to hear, and are used to hearing state-of-the-art leading-edge research. However: ASIC is an interdisciplinary conference and always has a diverse audience, Thus DO NOT give a talk aimed at your co-authors, laboratory colleagues, or even experts in your research domain: GIVE A TALK ACCESSIBLE TO AND UNDERSTANDABLE BY THE DIVERSE ASIC ATTENDEES.

Listing by speaker

[To be added later]

Speaker	Allen, Colin
Author 1	Allen, Colin University of Pittsburgh
Title	Probing the temporal structure of animal consciousness
Abstract	Temporal binding is the phenomenon in which events related as cause and effect are perceived by humans to be closer in time than they actually are. In a recent paper, Antonella Tramacere and I have proposed that this phenomenon can be investigated experimentally in nonhuman animals. In this talk I will outline our experimental design and discuss how the approach might be extended to study the temporal structure of animal consciousness more generally.

Speaker	Bhatia, Sudeep
Author 1	He, Lisheng Shanghai University
Author 2	Richie, Russell Children's Hospital of Philadelphia
Author 3	Bhatia, Sudeep University of Pennsylvania
Title	Limitations to Optimal Search in Naturalistic Active Learning
Abstract	We introduce a new empirical paradigm for studying naturalistic active learning, as well as new computational tools for jointly modeling algorithmic and rational theories of information search. Subjects in our task can ask questions and learn about hundreds of everyday items, but must retrieve queried items from memory. In order to maximize information gain, subjects need to retrieve sequences of dissimilar items. We find that subjects are not able to do this. Instead, associative memory mechanisms lead to the successive retrieval of similar items, an established memory effect known as semantic congruence. The extent of semantic congruence (and thus suboptimality) is unaffected by task instructions and incentives, though subjects are able to identify efficient query sequences when given a choice. Overall, our results indicate that subjects can distinguish between optimal and suboptimal search if explicitly asked to do so, but have difficulty implementing optimal search from memory. We conclude that associative memory processes place critical restrictions on people's ability to ask good questions in naturalistic active learning tasks.

Speaker	Breithaupt, Fritz
Author 1	Breithaupt, Fritz Indiana University, Cognitive Science & Germanic Studies
Title	The Narrative Mind
Abstract	Trough narratives we are able to reexperience past episodes and we are able to transform individual experience into shared experience. To achieve this, our minds and the ways in which we tell stories must be attuned to each other. But how exactly does this happen? The goal of the talk is to present the outlines of a model of narrative thinking that combines three aspects of narrative processing, namely the thinking in small episodes, multiversional processing, and narrative emotions. The talk will end by considering open questions.

Speaker	Breithaupt, Kira
Author 1	Breithaupt, Kira Indiana University
Author 2	Leite, Abe
Title	Analogy in behavior using Net-MATCH
Abstract	An understanding of analogy and the multiple realizability of concepts, ideas, and experience is necessary to understand cognition and the generation of behavior even at the most abstract levels. In this talk, I will (1) present a recently developed model of analogy-making called Net-MATCH, (2) give a high-level overview of the different domains it has successfully been applied to thus far, and (3) introduce several areas for future application of the model. Specific focus will be given to the latest application: assessing the analogical similarity between neural circuits evolved to generate the same behavior (a simple locomotion task) in different ways. In this domain, we found stronger analogies between circuits of different sizes that generate the behavior in similar ways than same-sized circuits that generate the behavior in different ways.

Speaker	Cottrell, Garrison
Author 1	Cottrell, Garrison UCSD
Author 2	Gahl, Martha UCSD
Author 3	Kulkarni, Shubham UCSD
Title	Modeling the primate visual system: An anatomically-constrained model
Abstract	For the last 40 years or so, I've been building computational models of cognitive processes. My model of face and object processing (The ModelTM) has explained a fairly large number of behavioral and fMRI results. Recently I have added more anatomically-constrained elements to a deep version, The Model 2.0. CNNs were inspired by the primate visual system (PVS), but there is still a large performance gap between human and machine vision. We believe it's time to take more features of the brain into account to bridge that gap. Our primary hypothesis is that the PVS was engineered by evolution for self-supervision, and hence should be optimally-structured for that purpose. We take four features of the PVS into account: 1) The foveated retina; 2) The log-polar mapping from retina to V1; 3) Active sampling using visual salience and attention; 4) The bifurcation of the ventral visual stream into central and peripheral pathways. Combining 2&3, samples on the same object give rise to very different images – natural data augmentation. The two pathways must agree in their representation of the visual world, like modern self-supervised learning systems (e.g., VICreg). Since the infant visual system is self-supervised, it seems quite possible that these architectural features are optimal for self-supervised learning.

Speaker	Cousineau, Denis
Author 1	Cousineau, Denis Université d'Ottawa
Title	Analyses of response time data in the Same-Different task
Abstract	The Same-Different task presents two stimuli in close succession and participants must indicate whether they are entirely the same or if some of the attributes between stimuli are different. While the task is simple and responses are typically very accurate, its results have proven difficult to explain. Notably, response times are characterized by a fast-same effect whereby Same responses are much faster than Different responses even though the Same stimuli must be exhaustively processed to be accurate. Herein, we examine a little more than a quarter million response times (N = 255,744) obtained from 327 participants who participated in one of 14 variants of the task involving minor variations in the stimuli or their durations. Response times are commonly summarized by averages; however, more information is provided by studying the entire distribution's characteristics, notably variance and skewness. Extending on the seminal work of Hockley (1984), we performed distribution fitting and analyzed estimated parameters stemming from the Ex-Gaussian and Weibull distributions. Covariances between the parameters and experimental manipulations were investigated to see whether they can be used to leverage process explanations of the task's results. These analyses exclude serial processing and do not support dual-route processing. The fast-same effect appears only through a shift of the response time distributions, a feature impossible to detect with mean response times analyses. Attention modulated process may be the most apt model of the fast-same effect.

Speaker	Cruz, Nicole
Author 1	Cruz, Nicole University of Innsbruck
Title	Measuring probabilistic coherence
Abstract	The probabilities we assign to statements, e.g. P(rain) = .6 and P(no-rain) = .4, are coherent whenever they follow the axioms of probability theory, so that if we used them in bets we would avoid a sure loss to one side. Coherence lies at the foundation of Bayesian inference. But how can we measure the extent to which people's inferences are coherent when, as is often the case, the information (or premises) on which the inferences are based is uncertain? To make such measurements, we must account for how logical constraints between probabilities may change when new information becomes available to reasoners; define and adjust for the probability of making a coherent response just by chance; and ascertain which patterns of premise probabilities would allow us to make plausibly falsifiable, and thus informative, assessments of sensitivity to coherence. I describe some of these challenges, and discuss how we might be able to tackle them in the quest to increase our understanding of reasoning under uncertainty.

Speaker	Dunn, John
Author 1	Dunn, John University of Western Australia
Author 2	Anderson, Laura Binghamton University
Title	Solving the problem of error in measurement theory: A procedure for testing the general linear model under monotonicity.
Abstract	Loftus (1978, Memory & Cognition 6, pp. 312-319) distinguished between a theoretical concept such as memory or attention, and its observed measure such as hit rate or percent correct. If the functional relationship between the concept and its measure is non-linear then only some interaction effects are interpretable. This is an example of the wider `problem of coordination' which pervades scientific measurement. Loftus drew on contributions to the representational theory of measurement (RTM) to discuss the consequences when the coordination function is assumed only to be monotonic. This led to the distinction between removable interactions that are consistent with an additive effect on the underlying theoretical concept and non-removable interactions that are not. However as noted by Wagenmakers, Krypotos, Criss and Iverson (2012, Memory & Cognition, 40, pp. 145-160), the adoption of these ideas by researchers has been greatly limited by the fact that no statistical procedure exists to determine if and to what extent an interaction is removable or otherwise. The lack of such a procedure has also been blamed for the negligible impact of RTM on research practice and has been listed as number two of 15 unsolved problems in the application of RTM. The aim of the present paper is to present such a procedure.

Speaker	Forbes, Eden
Author 1	Forbes, Eden Indiana University
Title	Ecosystem Thermodynamics & the Origins of Individuality
Abstract	Much work and biology and cognitive science have left us with a frustrating conclusion: most cognitively interesting organisms are themselves composites of other organisms that often themselves have cognitive features. This attribute of biological individuals leads to confusion about where the boundary of those individuals lies. Here, I hope to point to evidence from the thermodynamics of ecosystems for some scale-independent principles of individuality and how this field may help inform cognitive scientists.

Speaker	French, Bob
Author 1	French, Bob CNRS, France
Author 2	Defays, Daniel University of Liège, Belgium
Author 3	Tillmann, Barbara CNRS, Lyons Neuroscience Research Center
Title	Modeling early melody perception: a recurrent auto-encoder approach
Abstract	Are similar, or even identical, mechanisms used in speech segmentation, serial image processing and music perception? We address this question by exploring how TRACX2 (French et al., 2011; French & Cottrell, 2014; Mareschal & French, 2017), a recognition-based, recursive connectionist autoencoder model of chunking and sequence segmentation might be applied to elementary melody perception. The model, a three-layer autoencoder that recognizes "chunks" of short sequences of intervals that have been frequently encountered on input, is trained on the tone intervals of melodically simple French children's songs. It dynamically incorporates the internal representations of these chunks into new input. Its internal representations cluster in a manner that is consistent with "human-recognizable" melodic categories. TRACX2 is sensitive to both contour and proximity information in the musical chunks that it encounters in its input. It shows the "end-of-word" superiority effect demonstrated by Saffran et al. (1999) for short musical phrases. These results suggest that the recursive autoassociative chunking mechanism, as implemented in TRACX2, may be a general segmentation and chunking mechanism, underlying not only word- and image-chunking, but also elementary melody perception. An early version of this work was presented at ASIC 2019. Since that time, the model and simulations have been entirely reworked and the new results will be reported.

Speaker	Gedankien, Tamara
Author 1	Gedankien, Tamara Columbia University
Author 2	Tan, Ryan University of Texas Southwestern Medical Center
Author 3	Qasim, Salman Icahn School of Medicine at Mount Sinai
Author 4	Jacobs, Joshua Columbia University
Author 5	Lega, Bradley University of Texas Southwestern Medical Center
Title	Cholinergic modulation of hippocampal oscillations in humans
Abstract	Cholinergic synapses are essential for memory. While Alzheimer's disease and other age-related dementia entail loss of cholinergic pathways, therapies that promote cholinergic function deliver significant symptomatic benefits to patients. Yet, the neurophysiological mechanisms behind cholinergic modulation in the human brain remain unknown. Here, using rare intracranial recordings, we describe the effects of the anticholinergic drug scopolamine in the human medial temporal lobe (MTL) during a verbal memory task. We found that the memory impairment induced by scopolamine was associated with decreased low theta power during encoding as compared to placebo. Scopolamine also disrupted the phase reset of theta oscillations, and the magnitude of that disruption correlated with the degree of memory impairment induced by scopolamine. Lastly, we found that scopolamine significantly decreased interhemispheric synchrony in the theta band. Our findings show that the cholinergic system significantly influences the MTL oscillatory dynamics involved in successful encoding, and offers insights into the neural underpinnings of disorders affecting memory and learning in humans.

Speaker	Holden, Jay
Author 1	Holden, Jay University of Cincinnati, Center for Cognition, Action, and Perception
Author 2	Annand, Colin University of Cincinnati, Center for Cognition, Action, and Perception
Title	Holistic Approaches to Inference in Cognition and Action
Abstract	A non-reductive mode of system inference, derived from statistical physics, called the renormalization group, is explained. This coarse-graining framework is explained using randomness production performance. Binary randomness productions tasks challenge participants to use two buttons to produce binary sequences resembling random sequences, as if one button represented “heads” and the other “tails” in a long series of fair-coin flips. Studies dating back to the 1950's report biased sequences that alternate too often between options. These deviations from randomness are commonly modeled as emerging from limitations or idiosyncrasies in cognitive processing such as probabilistic biases and heuristics. A novel hypothesis that randomness departures are driven by coordination dynamics is tested. Fine-grained permutation analyses revealed the most common 1-5 trial sub-sequences were identified with the most dynamically stable coordinative relationships, consistent with bimanual coordination predictions. Coarse-grained power spectral density analyses revealed the sequences departed from classical randomness by virtue of membership in a more inclusive category of variability that subsumes classical randomness. Finally, a recurrence quantification analysis (RQA) quantified the intermediate scale dynamics of 5 to 25 trials. It revealed trade-offs between stochasticity and determinism in the participant's sequences that is characteristic of nonlinear dynamical systems. Renormalization group concepts suggest a close relationship between randomness production performance and two-alternative forced-choice performance, arising from the bimanual model's successful accounts of key performance patterns reported in both tasks.

Speaker	Ivanova, Iva
Author 1	Ivanova, Iva University of Texas at El Paso
Author 2	Uribe, Diana University of Texas at El Paso
Author 3	Hernandez, Dacia University of Texas at El Paso
Title	When we reuse interlocutors’ words: Does it make speaking easier, or harder?
Abstract	Speakers in a conversation sometimes reuse aspects of each other's language, a phenomenon known as entrainment. Entrainment facilitates comprehension, increases rapport and prosociality, and may underlie long-term learning and language change. In one prominent theory (Pickering & Garrod, 2004), entrainment is on most occasions automatic and therefore makes speaking itself cognitively easier. We examine the automaticity of entrainment to words in three picture-naming referential-communication experiments with bilinguals, relating entrainment magnitudes to speakers' language proficiency. Results isolate both automatic and non-automatic components of entrainment (less automatic with lower language proficiency), and show when entrainment does make speaking easier, and when it does not. The results have implications for clinical and educational settings.

Speaker	Janowski, Lucjan
Author 1	Janowski, Lucjan AGH University of Science and Technology
Author 2	Rusek, Krzysztof AGH University of Science and Technology
Author 3	Ćmiel, Bogdan AGH University of Science and Technology
Author 4	Nawała, Jakub Bristol University
Author 5	Wierzchoń, Michał Jagielonian University
Title	GSD - Modeling Lickert Scale with Discrete Distribution
Abstract	The Likert scale is a popular measure in psychology and related fields. However, there is no consensus how the data should be analyzed. In different studies such variables have be analyzed by simply ANOVA, a latent variable approach (e.g., ordered probit) or GLZ (Generalized Linear Models, ordinal regression). Typical solution is assuming a continuous distribution with eventual discretization. To avoid this, an appropriate discrete distribution class is required. We proposed a family of discrete probability distributions with only two parameters. The parameters can be easily interpreted, similar to normal distribution parameters. We call the new class the Generalised Score Distribution (GSD). The proposed GSD class covers the entire set of possible means and variances for any fixed and finite support, regardless of the number of answers. Furthermore, the GSD class can be treated as an underdispersed continuation of a reparametrized beta-binomial distribution. The parameters of the GSD class are intuitive and easy to estimate using the moment or MLE method. We are going to show how to use this model for psychological data.

Speaker	Katkov, Mikhail
Author 1	Katkov, Mikhail Weizmann Institute of Science
Author 2	Naim, Michelangelo McGovern Institute for Brain Research, MIT
Author 3	Tsodyks, Misha Weizmann Institute of Science, Institute for Advanced Studies
Title	Physicist view on memory recall
Abstract	Physicists would like to find a relationship between measured quantities that works in everyday life and are explained by simple principles, usually called basic, or fundamental laws. No such relationships in cognitive phenomena are known. We have recently proposed a deterministic memory retrieval process for the set of random items that leads to such relationship without free parameters, meaning it has universal applicability to any recall of random material, or a candidate for a fundamental law. The prediction of the model is verified in specially designed experiments to a surprising precision. In essence, the retrieval process is based on 3 assumptions: (1) the availability of memory for retrieval is binary it either can be retrieved in principle or not; (2) each item in memory is encoded in dedicated memory network by small fraction of neurons independently of other memories; (3) recall trajectory is determined by the matrix of encoding overlaps that we call 'similarity matrix' between items. These assumptions are sufficient to derive in the limit of sparse encoding a parameter free expression linking the number of memories available for recall and expected number of retrieved items. Both quantities were measured in experiments using recognition and recall experiment under the same presentation condition for the same group of participants and obtained good match to the predicted relationship. Furthermore, computing the second moment we were able to establish bounded region in a space spanned by first and second moments. This analysis may be performed when no recognition data available.

Speaker	Lewandowsky, Stephan
Author 1	Lewandowsky, Stephan University of Bristol
Author 2	Simchon, Almog University of Bristol
Author 3	Sutton, Adam University of Bristol
Author 4	Edwards, Matthew University of Bristol
Title	Psychological microtargeting in online environments
Abstract	In recent years, there has been a growing concern over "psychological microtargeting", in which psychological features that cannot be directly observed, such as personality characteristics, are inferred from online behavior and used for illegitimate purposes, for example to provoke political action/inaction or spread misinformation. Such microtargeting is opaque to online users and its potential effects remain unknown. Hence, there is an urgent need to reverse engineer microtargeting strategies by uncovering the algorithms in action. In this work, we present a proof of concept for such algorithmic reflection. Building on big data from the social media platform Reddit, we generated two models: (1) Predicting the personality makeup of users from produced text and (2) predicting the personality makeup of potential consumers of text. We recruited active Reddit users (N = 1100) of fiction-writing communities. The participants completed a Big Five personality questionnaire (BFI-2). They consented for their Reddit activity to be scraped and used in fitting a machine learning model. We trained a Bidirectional Encoder Representations from Transformers (BERT) model, predicting personality off produced text (average performance: r = 0.3x). We then applied Model 1 to a new set of Reddit users (N = 4000), predicted their personality based on their produced text, and trained a second BERT model to predict predicted-personality scores based on consumed text (average performance: r = 0.2x). By doing so, we provide the first glimpse into the linguistic markers of personality-congruent content. We discuss the implications of psychological microtargeting at scale, and how the detection of personality-congruent language can inform future interventions.

Speaker	Little, Daniel
Author 1	Little, Daniel The University of Melbourne
Author 2	Eidels, Ami The University of Newcastle
Title	Human scheduling of perceptual judgement and typing tasks
Abstract	Scheduling theory concerns the development of policies determining the optimal allocation of resources to a set of tasks. Scheduling problems have been studied extensively in the context of operations research and computer science, where optimal policies have been established for many cases, but almost no research has examined how people perform with respect to these optimal policies despite clear applications to selective attention and behavioral allocation. We conducted several experiments in which each subtask is a random dot motion (RDM) judgment, with difficulty determined by the coherence of the motion, or a typing task, with difficulty determined by the number of words required to be typed. Participants were presented with 4 or more subtasks, which are selected one at a time and then completed. We are concerned with the order in which subtasks are selected for completion. When subtasks vary in difficulty but have the same reward value, scheduling is optimal when tasks are completed from easiest to hardests. When reward also varies, then an optimal order index can be constructed from the ratio of reward to completion time. We compared a number of manipulations varying whether tasks appeared in a consistent or varied location, the deadline for completing the four subtasks, the number of subtasks, how subtask difficulty was labelled, and the nature of the reward (e.g., game-based or points-based). We introduce novel measures of performance and a model of scheduling allowing an analysis of optimality across conditions. In general, performance is more optimal or near optimal responding when rewards are equivalent and when location is fixed and there is a deadline.

Speaker	Ludwig, Kirk
Author 1	Ludwig, Kirk Indiana University, Bloomington
Title	What is Cognition? The Extended Mind Thesis and Group Cognition as Test Cases.
Abstract	The extent of cognition in the world depends on what it is. But to paraphrase Aristotle, “cognition” is used in many ways. The purpose of this talk is to generate some discussion about how “cognition” should be used in cognitive science. In broad strokes, we can use “cognition” liberally or strictly. What are the advantages of each? Using “cognition” liberally may capture commonalities across many different types of system, from colonies of bacteria to corporations, and to avoid the prejudice that cognitive processes end at the limits of the body so we can recognize how many problems we solve only using environmental aids. Using “cognition” more strictly may help us keep at arm’s length the dangers of projecting cognition onto or into systems when it lies only in the eye of the beholder and letting lazy thinking obscure relevant distinctions. As a foil for discussion, I advance a conservative thesis about cognition and apply it with negative effect to (1) the Extended Mind Thesis, according to which cognitive processes do not end at the limits of the body but extends into the environments through our use of tools in completing cognitive projects, and (2) the thesis of group cognition, according to which distributed problem solving (in teams) is group level cognition. The conservative thesis is that cognitive processes are sequences of states with original intentionality that conform to the standards of practical or theoretical rationality. I will show that this rules out both the extended mind thesis and group cognition.

Speaker	Melcher, David
Author 1	Melcher, David New York University Abu Dhabi
Author 2	Buonocore, Antimo Tuebingen University
Author 3	Huber-Huber, Christoph Donders Institute for Brain, Cognition and Behaviour; Radboud University
Author 4	Liu, Xiaoyi New York University Abu Dhabi
Title	The role of prediction in naturalistic visual perception: evidence from the extrafoveal preview effect
Abstract	We use saccadic eye movements to bring relevant information, glimpsed with extrafoveal vision, into the fovea for further processing. In contrast, most studies in vision science have used stable eye fixation and suddenly appearing and offsetting stimuli. This raises the question of how the extrafoveal preview of a saccade target influences visual object recognition in more natural viewing conditions. Research from my lab has shown that valid and predictable previews lead to faster and better face recognition judgments and reduced face-related evoked potentials. Similar effects are found with simple stimuli such as gratings. Overall, these studies suggest that visual perception during natural viewing is influenced by the extrafoveal preview, and prediction more generally. In line with theories of active, sensorimotor perception, studying behavioral and neural responses under more natural viewing conditions, in which the eye moves to objects of interest, may be necessary to understand how visual processing typically works in real life.

Speaker	Mermillod, Martial
Author 1	Mermillod, Martial Université Grenoble Alpes
Author 2	Mermillod, Martial
Title	Solving important limits of Deep Learning by returning to cognitive science
Abstract	The origins of Deep Learning and more generally of artificial neural networks are often unknown, including by current users of AI in computer science. I will recall the origins of Deep Learning, the current limits of these techniques which are unfortunately often used as magic wands, but also avenues for improvement by returning to psychology and cognitive neuroscience. I will present current results in the field of catastrophic forgetting with a wide range of applications for efficient, robust, frugal, realistic (as human, this psycho-inspired artificial neural network is agnostic and does not need a priori knowledge about the data) and plausible for edge AI (the model is bio-inspired and doesn't need neurogenesis or multi-heads) for life-long learning applications such as classification, facial identification or recognition of emotional expressions. I will also present promising lines of research avenues in the field of adversarial attacks, genesis of semantic information and motor control.

Speaker	Nelson Lowe, Angela
Author 1	Nelson Lowe, Angela University of CA, San Diego
Author 2	Wixted, John University of CA, San Diego
Title	Immediate Feedback Decreases False Recognition in the DRM Paradigm
Abstract	For some time, researchers have sought to gain a better understanding of the nature of false memories and the conditions that may decrease their likelihood of occurrence. The Deese Roediger McDermott (DRM) paradigm has been used extensively in this pursuit, utilizing study lists of words all semantically related to a theme word that is not actually presented (Deese, 1959; Roediger & McDermott, 1995). This design has been shown to be highly likely to produce a false recall or recognition of the theme word at test. Prior studies have aimed to combat this false recognition effect, with researchers providing warnings or feedback in the DRM task with varying degrees of success. Our study observed the effect of feedback in a DRM recognition task and found that participants who were given immediate correct/incorrect feedback after each memory judgment were less likely to make false recognitions to the theme word as the study progressed compared to those that did not receive feedback. The results suggest that feedback can decrease the likelihood of false recognition, possibly due to an increased reliance on memory features that are diagnostic of true memories.

Speaker	Pecher, Diane
Author 1	Pecher, Diane Erasmus University Rotterdam
Author 2	Zeelenberg, René Erasmus University Rotterdam
Title	The benefit of multisensory memory
Abstract	Is memory better for items that are studied as multimodal items than for items that are studied as unimodal items? Some researchers have found a benefit for multimodal items and have argued that this benefit is due to multisensory integration. An alternative explanation, however, is that more information is presented for multimodal than unimodal items, and that memory performance is positively related to the amount of information. We investigated the role of multisensory integration, replicating Thelen et al.’s (2015) method. Items were presented as unimodal (picture or sound) or bimodal (picture and sound) items in a continuous recognition task. In Experiments 1 to 4 we obtained no difference in memory performance between items studied as unimodal or bimodal stimuli, but there was a benefit of study-test overlap in format if sound was the task-relevant modality. We hypothesized that general multisensory benefits will be hard to obtain if the second modality is task-irrelevant. In Experiment 5 both picture and sound were task-relevant, and we found a benefit of bimodal study format. The benefit did not depend on temporal alignment of picture and sound, which rules out multisensory integration as an explanation. Instead, we argue that the memory benefit is due to the greater amount of information in bimodal rather than unimodal presentations.

Speaker	Plancher, Gaën
Author 1	Plancher, Gaën University of Lyon
Author 2	Purkart, Rudy University of Lyon
Author 3	Cavalli, Eddy University of Lyon
Author 4	Versace, Rémy University of Lyon
Title	No sensory visual imagery in aphantasia?
Abstract	For most people mental images are omnipresent in daily life. But for some people the experience of visual imagery without direct external stimulus is lacking, a condition recently referred as to aphantasia. So far, most of the studies on aphantasia rely on subjective reports, since participants are required to generate mental images and/or assess the vividness of these mental images. In the present study, the formation of mental images was estimated in individuals with aphantasia without explicitly asking them to generate mental images. Participants performed an implicit priming task where a probe is assumed to automatically reactivate a mental image. An explicit priming task, where participants were explicitly required to form a mental image after a probe, served as a control task. While control participants showed a priming effect for the implicit and the explicit tasks, participants with aphantasia did not show any priming effects. These results may suggest that aphantasia relies on a genuine inability to generate mental images rather than to a deficit in accessing these images. Interestingly, when we put all the participants together and considered the score at the Vividness of Visual Imagery Questionnaire as a continuous variable, it correlated with the size of the implicit priming effect, but not with the explicit one, suggesting that our implicit measure might be relevant to screen aphantasia.

Speaker	Richerme, Phil
Author 1	Richerme, Phil Indiana University
Title	Making Heads and Tails of Quantum Information Science
Abstract	The field of quantum information promises a technological revolution, wherein quantum-based devices will push beyond the computing limits established by classical physics and surpass the capabilities of the world's best supercomputers. If successful, its impact will span diverse fields such as computer science, artificial intelligence, machine learning, systems modeling, drug design, and information security. Yet, “quantum” is an often overused buzzword that is over-hyped in popular media and is often applied incorrectly. In this talk, I will describe the fundamental aspects of systems which exhibit quantum behavior and what truly distinguishes them from classical systems. Next, I will show examples of well-controlled quantum systems (including those from my own lab) which can be used to generate non-classical effects and calculate behaviors inaccessible to classical computation. Finally, I will provide an outlook for quantum information processing and discuss the most likely near- and far-term applications.

Speaker	Roskies, Adina
Author 1	Roskies, Adina Dartmouth College
Title	What is the Readiness Potential?
Abstract	Neuroscientific results have sometimes been claimed to demonstrate that we lack free will. Perhaps the best known of these results are the Libet experiments, which claim to show that our brains decide to initiate actions before we consciously decide to do so. Recent work in computational modeling, and recent experiments exploring the readiness potential paint a different picture. I will argue that the Libet interpretation of experiments using the RP is a misinterpretation of the data, and that putting the experiment in context reveals a picture consistent with commonsense views of free will. I explore the implications for this alternative interpretation for the philosophical problem of free will.

Speaker	Sauciuc, Gabriela-Alina
Author 1	Sauciuc, Gabriela-Alina Lund University, Cognitive Science
Author 2	Persson, Tomas Lund University, Cognitive Science
Title	Where does human cooperation come from? The evolutionary origins of the ability to infer shared goals and motivations
Abstract	Humans have an irresistible inclination to coordinate actions with others, leading to species-unique forms of cooperation. According to current theories, human cooperation is made possible by shared intentionality (henceforth SI), typically defined as a motivation to share psychological states with others, which in turn enables individuals to act jointly in the mutually aware pursuit of shared goals. SI is hypothesized to have evolved as late as 400 000 years ago, as our ancestors (in particular Homo heidelbergensis), turned to a kind of food procurement that required joint coordinated action. Thus, SI is hypothesized to be absent in other species, including our closest genetic relatives, the nonhuman great apes (henceforth apes), whose psychology is claimed to be entirely driven by individualistic motivations. This seeming divide between human and ape psychology entails an absence of precursors traits to investigate in non-human species, which hampers the comparative and evolutionary study of SI. In this presentation, we will provide an overview of recent empirical findings - including findings from our own previous and ongoing research - that contradicts the hypothesis that SI is absent in apes. This evidence shows that traits regarded as constitutive of SI (such as joint attention, commitment, referential communication, action co-representation, re-engagement in social games, etc.) are present in apes. In turn, this is consistent with the alternative hypothesis that SI has ancient evolutionary roots, and some of its constitutive traits were already present in the last common ancestor of the great apes (Hominidae), prior to the split between Ponginae (orangutans) and Homininae (gorillas, chimpanzees, humans).

Speaker	Servant, Mathieu
Author 1	Servant, Mathieu University of Franche-Comté (France)
Author 2	Logan, Gordon Vanderbilt University (USA)
Author 3	Gajdos, Thibault Aix-Marseille University (France)
Author 4	Evans, Nathan University of Queensland (Australia)
Title	An integrated theory of deciding and acting
Abstract	"One would expect psychology—the science of mental life and behavior—to place great emphasis on the means by which mental life is behaviorally expressed. Surprisingly, however, the study of how decisions are enacted—the focus of motor control research— has received little attention in psychology." (Rosenbaum, 2004, American Psychologist). Our integrated theory of deciding and acting (Servant, Logan, Gajdos, & Evans, 2021, JEP General) assumes that motor execution in choice tasks is determined by the same evidence accumulation variable that drives decision-making. The theory builds upon recent insights from the neuroscience of decision-making and motor control. It is formalized as an extension of Ratcliff 's diffusion model, and assumes that two thresholds operate on the evidence accumulation decision variable. The ﬁrst threshold, referred to as electromyographic (EMG) threshold, marks the onset of electrical activity in the response-relevant muscle and the beginning of force production. The second threshold corresponds to the response. I'll review empirical evidence for the model, and discuss current challenges.

Speaker	Shiffrin, Richard
Author 1	Shiffrin, Richard Indiana University
Author 2	Macleod, Colin Vanderbilt University
Author 3	Maxcey, Ashleigh Vanderbilt University
Author 4	Nosofsky, Robert Indiana University
Author 5	Cutler, Rebecca Vanderbilt University
Author 6	Joykutty, Zara Vanderbilt University
Title	What are the causes of forgetting?
Abstract	What are the chief causes of forgetting? Several causes are well established and accepted: For short-term memory, limited capacity is a cause of forgetting. There are reasons to believe long-term memory does not suffer from a similar capacity limitation and that memory traces remain more or less permanent if undisturbed. Nonetheless those traces may exhibit forgetting due to failures of retrieval. Accepted causes of retrieval failure are competition (aka interference, cue overload), context change, and addition to traces that occur when they are retrieved.. Bob Bjork, Elizabeth Bjork and Michael Anderson highlighted another possible cause of long-term forgetting in a 1994 paper reporting a cued recall paradigm that exhibited Retrieval Induced Forgetting (RIF). They argued that the RIF was due to inhibition causing a reduction in trace strength of a trace that is retrieved but is not the one desired. Ashleigh Maxcey and colleagues published several papers using a design like that of Bjork et al., but using recognition rather than cued recall. They also found RIF, and also interpreted that as due to inhibition. We report two tests of inhibition, one in the recognition paradigm and one in the cued recall paradigm, finding no RIF in either, showing that inhibition does not cause forgetting or is at most a minor contributor to forgetting.

Speaker	Speekenbrink, Maarten
Author 1	Speekenbrink, Maarten Experimental Psychology, University College London
Title	Trace decay in stigmeric processes
Abstract	Stigmergy refers to a mechanism of indirect coordination, where an action by one agent leaves a trace in the shared environment that stimulates an action of a successive agent. By contrast to direct communication and observation, stigmeric processes don't require the acting and observing agent to be present in the same location and time. Behavioural traces are also devoid of contextual information about the acting agent, so that behavioural traces can extend beyond social boundaries. As such, behavioural traces are potentially a powerful mechanism for behaviour change. In this presentation, I will provide a relatively simple mathematical analysis of stigmeric processes - based on Markov chains - with a particular focus on how the decay rate of behavioural traces affects performance in the population of agents, when the environment is and is not perturbed by unsignalled changes in what can be deemed optimal actions. This work is inspired by a study by Topf & Speekenbrink (2021, https://psyarxiv.com/496v7/) who investigated recycling decisions in controlled lab experiments and in an intervention study manipulating and enhancing the visibility of prior recycling decisions.

Speaker	Testolin, Alberto
Author 1	Testolin, Alberto University of Padova
Author 2	Cognolato, Samuel University of Padova
Title	Transformers discover an elementary calculation system exploiting local attention and grid-like problem representation
Abstract	Mathematical reasoning is one of the most impressive achievements of human intellect but remains a formidable challenge for artificial intelligence systems. In this work we explore whether modern deep learning architectures can learn to solve a symbolic addition task by discovering effective arithmetic procedures. Although the problem might seem trivial at first glance, generalizing arithmetic knowledge to operations involving a higher number of terms, possibly composed by longer sequences of digits, has proven extremely challenging for neural networks. We show that universal transformers equipped with local attention and adaptive halting mechanisms can learn to exploit an external, grid-like memory to carry out multi-digit addition. The proposed model achieves remarkable accuracy even when tested with problems requiring extrapolation outside the training distribution; most notably, it does so by discovering human-like calculation strategies such as place value alignment.

Speaker	Thalmann, Mirko
Author 1	Thalmann, Mirko Max Planck Institute for Biological Cybernetics
Author 2	Schaefer, Theo Max Planck Institute for Human Cognitive and Brain Sciences
Author 3	Theves, Stephanie Max Planck Institute for Human Cognitive and Brain Sciences
Author 4	Doeller, Christian Max Planck Institute for Human Cognitive and Brain Sciences
Author 5	Schulz, Eric Max Planck Institute for Biological Cybernetics
Title	Are Representations Shaped by Task-Specific Goals?
Abstract	Do mental representations change in an adaptive way? Broadly, previous theories assume that each encounter with a stimulus leads to a separate episode being stored or that encoding the same stimulus several times leads to a more complete and perhaps less error-prone representation of the stimulus. In contrast, we suggest a theory that predicts that representations change in an adaptive way. The theory assumes that a sensory representation is a noisy version of a presented stimulus and only stored in memory if it is helpful to achieve the current task goal. We test qualitative predictions of that theory in an experiment in which participants learn to categorize two-dimensional stimuli into two categories. In short, the theory predicts that representations are attracted by the category centers and repulsed by the category boundaries. We measure participants’ representations in a continuous-reproduction task before the categorization task as a baseline and after the categorization task to examine representational change. The data collected so far deviate from the qualitative predictions. We discuss issues with regards to the experimental design and potential changes to it to further test the theory.

Speaker	Wierzchon, Michal
Author 1	Wierzchon, Michal Jagiellonian University
Author 2	Rutiku, Renate Jagiellonian Univeristy
Author 3	Nitka, Aleksander Jagiellonian Univeristy
Author 4	Sandberg, Kristian Aarhus University
Title	The neural architecture of consciousness - database and some primary analyses
Abstract	Despite years of research investigating Neural Correlates of Consciousness (NCC), researchers in the field still have not reached an agreement on the nature of the involved mechanisms and their localisation. Two contradictory views are vividly discussed in the literature. According to the first, consciousness is related to the relatively late stage of information processing, and thus the NCC should be found in the frontal cortex. The second claims that consciousness is related to the early stages of processing and favours involvement of the posterior parts of the brain. Our central hypothesis is that neither of those views can be sufficiently justified, as they strongly depend on the conceptualisation of what is measured, operationalisation of the consciousness, and the kind of task used. We claim that such an approach misses the bigger picture and that consciousness may require early activation of the sensory system and the late activation of the prefrontal cortex depending on the task condition and applied operationalisation of consciousness. Recently, we have proposed an alternative approach to this problem, focusing on individual differences in experience and looking for NCCs that are not content dependent. In order to comprehensively investigate individual differences in consciousness, we have collected a variety of behavioural tasks and combined them with the structural and functional brain’s organisation as measured with magnetic resonance imaging (MRI). I will present the collected database and the preliminary analyses aimed at investigating individual differences in consciousness and searching for the NCC in a complex and multidimensional manner.

Contact reberle@indiana.edu with questions.