Seventh Annual Summer Interdisciplinary Conference

Authors, Titles, Abstracts


(When sufficient titles, and abstracts for talks and posters arrive, I will begin posting them in this section)

Alphabetical listing by speaker

A | B | C | D | E | F | G | H | I | J | K | L | M | N | O
P | Q | R | S | T | U | V | W | X | Y | Z

B
Speaker's Name:Bruno Bocanegra
First Author's Name:Bruno Bocanegra
First Author's Affiliation:Erasmus University Rotterdam
Second Author's Name:René Zeelenberg
Second Author's Affiliation:Erasmus University Rotterdam
Title:Emotion impairs high frequency spatial vision
Abstract:In order to respond adaptively to threat in the environment our brains are equipped with specialized mechanisms that enhance the visual processing of emotional events. Although a previous study indicated that emotion enhances vision, the generality of this finding remains unknown. Do the benefits of emotion extend to all basic dimensions of vision or are they limited in scope? Here, we provide the first demonstration that emotion not only improves but also impairs low-level vision. Our results indicate that the brief presentation of a fearful face enhances orientation sensitivity for low spatial frequency gabors, but diminishes sensitivity for high spatial frequency gabors. We show this counterintuitive pattern of benefits and deficits is due to a sensitivity shift across the spatial frequency spectrum, triggered by the global facial configuration in the fearful expressions. Consistent with previous neuroimaging data, the selective low-frequency benefits suggest that emotion enhances magnocellular visual processing. Additionally, we propose that the high-frequency deficits might be due to cross-inhibition between magno- and parvocellular visual pathways. Our results reveal an emotional mechanism that improves the detection of coarse features at the expense of fine-grained visual details, presumably in order to facilitate responses to motivationally significant stimuli.

C
Speaker's Name:Andrew Cohen
First Author's Name:Andrew Cohen
First Author's Affiliation:University of Massachusetts
Second Author's Name:Michael Ross
Title:Exploring mass ratio perception with psychological Markov chain Monte Carlo sampling
Abstract:A number of previous studies have examined humans' ability to deduce the relative mass of objects in idealized collisions. Using a newly developed technique of psychological Markov chain Monte Carlo sampling (Sanborn & Griffiths, 2006), this work examines subjects' mental representations of different collision mass ratios. The results reveal strong inter-subject similarities and qualitative differences between the representations of 1:1 and 1:2 ratios. The results suggest that while subjects have a narrow and constrained (albeit incorrect) internal model of 1:1 collisions, their representations of other ratios are vague and diffuse.

D
Speaker's Name:Eddy Davelaar
First Author's Name:Eddy Davelaar
First Author's Affiliation:Birkbeck, University of London
Title:Primary memory: contributions to free recall performance
Abstract:On the same page on which William James introduced the terms primary and secondary memory, he presented his view on how they are implemented in the brain, which is not different from Hebb's view. Here I start with the view that primary memory is the activated part of long-term memory, or better yet that primary memory is the process (see also, Norman, 1968) by which temporarily activated long-term knowledge is maintained in active state beyond its expected unaided life-time. The current debate in the literature is whether a short-term buffer needs to be postulated in order to account for performance on free recall tasks (Brown, et al, 2007; Davelaar, et al, 2005; Howard, et al, 2007). I will briefly touch on dissociations in recency effects over the short- and long-term, on dissociations between immediate and longer-term free recall tasks, and some points of confusions. One source of confusion is the unclarity of the concept of activation-based short-term store. I will show using simulations why an activation-buffer is different than a fixed-box buffer and its association with working memory models that assume that the content of working memory is the activated part of long-term memory (Cowan, 1999; McElree, 2006; Oberauer, 2002).

F
Speaker's Name:Stefan Frank
First Author's Name:Stefan Frank
First Author's Affiliation:University of Amsterdam
Title:Testing the Surprisal Theory of Word-reading Time
Abstract:According to so-called "surprisal theory" (Hale, 2001; Levy, 2008), a word's probability of occurrence given its sentence context is inversely logarithmically related to the time required to read that word. Tests of this theory have implicitly assumed that the (subjective) probabilities a reader assigns to words correspond to the (objective) probabilities as extracted from text corpora. If surprisal theory is correct and subjective probabilities indeed correspond to objective probabilities, an objectively more accurate probability model should also provide more accurate predictions of word-reading times. To investigate whether this relation holds, we compared two models that can generate word probabilities and that have been suggested as a basis for psycholinguistic models of sentence processing. However, they make very different assumptions as they originate from disparate fields: One is a Simple Recurrent Network (SRN; the quintessential connectionist model), the other a Probabilistic Context-Free Grammar (PCFG; the standard model in computational linguistics). Both models were trained on a part of the Wall Street Journal corpus and tested on the Dundee corpus, which contains both newspaper texts and corresponding eye-movement data. Preliminary results show that the SRN generates more accurate word probabilities, whereas the PCFG provides better predictions of reading times. This suggests that subjective probabilities cannot be estimated by objective probabilities or that surprisal theory is simply incorrect. If we do hold on to the theory, the results indicate that a system based on tree-structures forms a better psycholinguistic model than does a connectionist system.

Speaker's Name:Bob French
First Author's Name:Bob French
First Author's Affiliation:LEAD-CNRS, U. of Burgundy, Dijon, France
Second Author's Name:Cowell Rosemary
Second Author's Affiliation:LEAD-CNRS, U. of Burgundy, Dijon, France
Title:Rule Extraction in Category Learning: a semi-supervised neural network model
Abstract:We develop a semi-supervised “dual-network” connectionist model of category learning in which rules gradually emerge from a standard Kohonen network. The architecture is based on the interaction of a statistical-learning (Kohonen) network and a competitive-learning rule network. The rules that emerge in the rule network are essentially weightings of individual features according to their importance for categorisation. These rules emerge due to the presence of noise on the input of the Kohonen network. Once the combined system has learned a particular rule, it de-emphasizes those features that are not sufficient for categorisation, thus allowing correct classification of novel, but atypical, stimuli, for which a standard Kohonen network fails. We will explain the principles and architectural details of the model and show how it works correctly for stimuli that are misclassified by a standard Kohonen network.

H
Speaker's Name:David Huber
First Author's Name:Yoonhee Jang
First Author's Affiliation:University of California, San Diego
Second Author's Name:David Huber
Second Author's Affiliation:University of California, San Diego
Title:Measuring the statistical dependence between familiarity and recall
Abstract:Following study of paired words, participants were given a sequence of test trials that alternated between 1) forced choice recognition of single words and 2) use of the target from step 1 as a cue for recall of the studied associate. This recognition followed by cued recall procedure produces a 2X2 table for memories that do or do not support recognition combined with memories that do or do not support cued recall. In addition, recognition performance was manipulated through immediate repetition priming (i.e., the Jacoby-Whitehouse paradigm) and also through speeded versus non-speeded responding. Multinomial Processing Tree (MPT) models of these data were used to examine the influence of priming on recognition and the influence of speeded responding on recognition. Finally, an appropriate MPT was used to measure the statistical dependence between familiarity based recognition versus recall based recognition.

K
Speaker's Name:George Kachergis
First Author's Name:George Kachergis
First Author's Affiliation:Indiana University
Second Author's Name:Richard Shiffrin
Second Author's Affiliation:Indiana University
Title:Context Effects in Recognition Memory: A Bayesian Analysis
Abstract:Many people have had the experience of knowing what song will play next on an album (even one heard only a few times). Conversely, many people fail to recognize an acquaintance encountered in an unfamiliar context. Our studies explore the automatic or incidental storage of associations between successively encountered words on a list (storage of such associations is a key assumption found in the models of Howard and Kahana, and the REM-II model of Mueller, Nelson, and Shiffrin). We find evidence in episodic recognition memory for such storage. We explore the effects of the previous item’s familiarity, semantics, and study-list position on the accuracy and response time for a target word preceded at test by same, similar, or different context words. We use Hierarchical Bayesian data analysis to motivate the formulation of a model.

Speaker's Name:Krystal Klein
First Author's Name:Krystal Klein
First Author's Affiliation:Indiana University
Second Author's Name:Chen Yu
Third Author's Name:Richard Shiffrin
Third Author's Affiliation:Indiana University
Title:What You See is What You Learn
Abstract:Recent studies suggest that both adults and infants can potentially solve the reference uncertainty problem in language learning through the use of cross-situational statistical information (e.g. Yu & Smith, 2007). Here we present data from several new experiments wherein adults are exposed to a rapid series of learning trials containing uncertainty in sound-to-picture mappings, but in which this uncertainty is resolved across multiple trials. By monitoring the eye movements of adult participants with various degrees of initial knowledge about sound-to-picture mappings, we provide added constraints upon models of temporally-extended statistical language learning (Yu, Smith, Klein & Shiffrin, 2007).

Speaker's Name:Maria Kozhevnikov
First Author's Name:Maria Kozhevnikov
First Author's Affiliation:George Mason University
Title:The role of immersivity in three-dimensional spatial transformations
Abstract:The majority of experimental studies on 3D visual-spatial processing have been conducted using traditional 2D displays. We were interested in the contribution of immersion to 3D image transformations and compared subjects’ performance on spatial transformation tasks within traditional 2D, 3D non-immersive (stereo glasses), and 3D-immersive (head mounted display with position tracking) environments. Fifteen participants completed perspective-taking test where they imagined transforming their own perspective to another perspective in a computerized scene and pointed to one of several objects from the new perspective. In addition, the participants completed a Shepard and Metzler Mental Rotation Task, in which they were asked to mentally rotate 3D objects along the picture (X), vertical (Y), or depth (Z) axes. While the patterns of subjects’ responses were not significantly different in 2D and 3Dnon-immersive environment, we found a unique pattern of responses in 3D immersive environment, suggesting that immersion triggered significantly greater use of body-related egocentric object coding and visuo-motor strategies than two other non-immersive environments. In particular, while the subjects used scene-based spatial frame of reference in non-immersive environments, they use viewer-center encoding and body-related frame of reference while performing spatial transformation tasks in a 3D immersive environment. Overall, our findings suggest that 3DI environments are different from 3D non-immersive and 2D traditional environments, and that only 3D immersive environments provide adequate information for building the spatial reference frame crucial for high-order motor planning and egocentric encoding.

L
Speaker's Name:Daniel Lafond
First Author's Name:Daniel Lafond
First Author's Affiliation:Université Laval
Second Author's Name:Yves Lacouture
Second Author's Affiliation:Université Laval
Title:Testing the predictive accuracy of decision tree models of categorization
Abstract:This work examines the descriptive and predictive accuracy of three decision tree models of categorization adapted from Trabasso, Rollins and Shaughnessy (1971). These models aim to provide a quantitative account of categorization response times, choice proportions and typicality judgments at the individual-participant level. Study I modeled results from Cohen and Nosofsky’s (2003) experiment. Overall, the decision tree models achieved comparable fits to that of two exemplar models, the EGCM-RT (Lamberts, 2000) and the EBRW-PE (Cohen & Nosofsky, 2003). In Study II, we replicated and extended Cohen and Nosofsky's experiment by asking participants to give subjective typicality ratings for each stimulus. A post-test phase called the “four-questions game” (Sayeki, 1969) provided the constraints required to systematically identify a unique decision tree for each participant. Model I, II and III showed increasingly good fits to the data, which follows from their respective complexity and flexibility. However, there is a risk that more flexible models provide better fits simply by adjusting to noise in the data (overfitting). We compared the predictive accuracy of Model I, II and III using a cross-validation procedure. The predictive accuracy of the three models was generally good, though a different degree of overfitting was observed depending on the model, on the dependent variable and on the number of trials in the test phase of the experiment (i.e., the level of noise). This test provides a challenging new benchmark for comparing categorization models.

Speaker's Name:Hsin-I Liu
First Author's Name:Hsin-I Liu
First Author's Affiliation:University of the Incarnate Word
Title:Communication as a disguise of symbolic violence: Bourdieu’s structural analysis of socio-cultural domination
Abstract:This paper will argue that the most important theoretical question for Bourdieu is how symbolic violence is reproduced and transmitted in contemporary capitalist societies. Symbolic violence can be briefly defined as “the form of domination that is only exerted through the communication in which it is disguised.” In other words, Bourdieu provocatively claims that the relations of communication are always interwoven with relations of power, and that no human communication is without socio-cultural domination. Such a domination is structured in capitalist societies and insured not by ideological control but by institutional mechanisms—especially those on "the reinforcement of predispositions." Moreover, Bourdieu indicates that the mode of domination is always inscribed in communication. “The whole content of communication (and not just the language used),” as he eloquently puts it, “is unconsciously modified by the structure of the relationship between speakers." However, in my view, the major theoretical difficulty facing Bourdieu is how the dominant class—those who have material and symbolic capitals—ensures that the members of society (especially the dominated class) accept or reproduce their discourses of social reality, i.e. symbolic violence, when different interpretants have different interpretations and the process of interpreting is continuously changing. The ultimate question becomes: how does the dominant group guarantee individual dissensions and practical tactics to be confined to minimum and local levels that totally suppress major forms of challenge or threat to this dominating structure of symbolic violence in capitalist societies?

Speaker's Name:Casimir Ludwig
First Author's Name:Casimir Ludwig
First Author's Affiliation:University of Bristol
Title:Temporal integration underlying saccadic eye movement decisions
Abstract:Saccadic eye movements are among the most frequent perceptual decisions humans make. In recent years more general models of perceptual decision making, such as signal detection and sequential sampling models, have been applied to account for where observers look (and – in the case of sequential sampling models – when they look there). I will review applications of these models in the oculomotor domain, with particular emphasis on models that assume sensory evidence in favour of a number of potential saccade targets is integrated over time up to some criterion level. Next, I will present data from a number of studies that illustrate this strategy does not always hold. When observers are required to select a saccade target on the basis of luminance contrast in the presence of external noise, their eye movement decisions appear to be based on a very limited and fixed epoch of sensory evidence. This epoch is not adjusted according to the strenght of this evidence. Finally, I will present a number of alternative models and assess their ability to account for the experimental data. These models allow for inferences to be made regarding 1) the nature of the temporal filter that provides the input to a downstream "decision unit"; 2) whether the decision unit integrates the signal over time or compares momentary values against some threshold; 3) the duration of the integration/sampling window.

M
Speaker's Name:Michael McKenna
First Author's Name:Michael McKenna
First Author's Affiliation:Florida State University, Philosophy
Title:Reasons-responsiveness and free will
Abstract:I identify free will with all of the control required for moral responsibility, and I analyze that control in terms of an agent's responsiveness to reasons. The puzzle is how to capture the proper spectrum of responsiveness as required for morally responsible agency. Although I follow Fischer and Ravizza's 'mechanism-based' approach, unlike them, I do not remain agnostic about how to individuate the mechanisms. In my presentation, I offer a proposal for mechanism individuation.

N
Speaker's Name:Angela B. Nelson
First Author's Name:Angela B. Nelson
First Author's Affiliation:Indiana University
Second Author's Name:Richard M. Shiffrin
Second Author's Affiliation:Indiana University
Title:How Experience Shapes Memory
Abstract:Novel items (Chinese characters) were trained in a visual search task to have differential experience. The induced frequency differences were shown to produce marked effects on episodic and implicit memory transfer tasks (Nelson & Shiffrin, 2006). Because each novel item is randomly assigned for each participant to a frequency category, these results are inconsistent with the REM model account of frequency effects (Shiffrin & Steyvers, 1997). This account posited higher frequency items to have higher frequency features. We present a new variant of the REM model incorporating contextual diversity in an item’s representation, particularly the diversity imposed by different item sets simultaneously in short term memory during training: Features of such nearby items join a target item’s representation. This model is a simplification of the REM-M model proposed by Mueller and Shiffrin (e.g. 2006), and is related to the ideas found in the TCM model of Howard and Kahana (e.g. 2002). Because the higher frequency items are seen in a larger variety of contexts than lower frequency items, the higher frequency items develop a more diverse representation in the lexicon. The model is shown to account for frequency effects found by Nelson and Shiffrin (2006).

O
Speaker's Name:Chrispine OMONDI
First Author's Name:Chrispine OMONDI
First Author's Affiliation:Coffee Research Foundation, P O Box 4, 00232 Ruiru, Kenya.
Title:COFFEE QUALITY ASSESSMENT: THE CASE OF TWO KENYAN CULTIVARS, RUIRU 11 AND SL 28
Abstract:It is widely recognized that coffee quality is the single most important factor that determines coffee prices in the international market. The prominence given to Kenya coffee world-wide is derived from the fine quality coffee it supplies to the world market. Therefore selection for desirable attributes of coffee quality has been the subject of detailed studies in the development of coffee varieties in Kenya. Quality parameters of two Kenya coffee cultivars Ruiru 11 and SL 28 believed to be similar in their major quality attributes were assessed to determine factors that influence the final beverage quality. Both varieties were grown in two locations in different agro-ecological zones. The parameters assessed were bean quality, beverage quality and overall class. The data was subjected to multivariate hierarchical cluster analysis based on Eucledian Distance Matrix Method. The study revealed that the growing environment had a strong effect on the expression of quality parameters exhibited by Ruiru 11 and SL 28. The study also tested the consistency of cup tasters to evaluate similar samples and arrive at similar results. It was revealed that there was significant difference among cup-tasters. In other words, when cup tasters were presented with similar blind samples for assessment, the results varied from one cup-taster to the other. The study recommends proper management of the growing environment to obtain the desired quality attributes and branding to meet the preferred taste of different consumers.

Speaker's Name:Chrispine OMONDI
First Author's Name:Chrispine OMONDI
First Author's Affiliation:Coffee Research Foundation, P O Box 4, 00232 Ruiru, Kenya.
Title:COFFEE QUALITY ASSESSMENT: THE CASE OF TWO KENYAN CULTIVARS, RUIRU 11 AND SL 28
Abstract:It is widely recognized that coffee quality is the single most important factor that determines coffee prices in the international market. The prominence given to Kenya coffee world-wide is derived from the fine quality coffee it supplies to the world market. Therefore selection for desirable attributes of coffee quality has been the subject of detailed studies in the development of coffee varieties in Kenya. Quality parameters of two Kenya coffee cultivars Ruiru 11 and SL 28 believed to be similar in their major quality attributes were assessed to determine factors that influence the final beverage quality. Both varieties were grown in two locations in different agro-ecological zones. The parameters assessed were bean quality, beverage quality and overall class. The data was subjected to multivariate hierarchical cluster analysis based on Eucledian Distance Matrix Method. The study revealed that the growing environment had a strong effect on the expression of quality parameters exhibited by Ruiru 11 and SL 28. The study also tested the consistency of cup tasters to evaluate similar samples and arrive at similar results. It was revealed that there was significant difference among cup-tasters. In other words, when cup tasters were presented with similar blind samples for assessment, the results varied from one cup-taster to the other. The study recommends proper management of the growing environment to obtain the desired quality attributes and branding to meet the preferred taste of different consumers.

R
Speaker's Name:Adina Roskies
First Author's Name:Adina Roskies
First Author's Affiliation:Dartmouth College
Title:What can neuroscience tell us about free will
Abstract:I will discuss the problem of free will in the context of new information from neuroscience. Specifically, I will consider whether the picture of decision-making that is emerging from neuroscience provides reason to think our actions are determined or not under our control. I provide several reasons to think that our philosophical positions about the correct relation between freedom and determinism is immune to input from the biological sciences, but I also argue that dependency on determinism is misconceived. Instead, I outline another view, and show how our philosophical views about free action can be consistent with a mechanistic model of brain function, and illustrate this by reference to current neuroscientific work.

S
Speaker's Name:Jean-Christophe Sarrazin
First Author's Name:Jean-Christophe Sarrazin
First Author's Affiliation:UMR CNRS Mouvement et Perception
Second Author's Name:Emmanuel Daucé
Second Author's Affiliation:UMR CNRS Mouvement et Perception
Third Author's Name:Axel Cleeremans
Third Author's Affiliation:Université Libre de Bruxelles
Title:Psychological hysteresis in sudden awareness: a functionnal correlate of recurrent processing
Abstract:In this study we explore the idea that sudden motor awareness emerges as the result of global competition biased by top-down modulation, which implements global constraint satisfaction. The contents of conscious experience at some point in time thus reflect the application of the brain’s knowledge on the current situation so as to yield the most adapted representations in the service of action. Such processes in turn critically depend on recurrent, or reentrant processing. Subjects carried out a reaching movement, which was disturbed by a haptic arm on some trials. Participants (1) pointed at the target as accurately as possible before returning to the start position, making a visuomotor adjustment to the target if required and (2) reproduced the spatial path of the movement they had just made, as accurately as possible, to give an indication of their awareness of the pointing movement. We analysed the spatial disparity between the initial and the reproduced movements on those with a movement disturbance. In this framework, we assume that once a recurrent neural activity due to the disturbance has reached a certain threshold, the visuomotor adjustment is suddenly consciously perceived so that the conscious activity may show hysteresis---as Libet (1973) has claimed---since the activity is probably held above a threshold to some extent by some mechanism, such as loops with positive feedback. This activity corresponds to the neural correlates of consciousness (NCC). Introducing a minimalist dynamics model, we highlight the computational principles of the conscious perception process.

Speaker's Name:Lawrence Shapiro
First Author's Name:Lawrence Shapiro
First Author's Affiliation:University of Wisconsin - Madison
Title:Representation and the Philosophy of Science
Abstract:Embodied, Embedded, and Situated Cognitivists have in recent years offered accounts of cognition that seek to minimize or modify traditional conceptions of representation in cognitive explanation (see, e.g. Beer 1993, Thelen xxxx, Brooks 1991). Some of these efforts draw on notions like simplicity, unification, and indispensability that philosophers of science have studied for decades. However, it is not clear how concepts like these "fit" into theories of cognition (how are they to be measured?), and so appeals to these notions in order to justify a particular stand on representation are hard to evaluate. In this paper I examine some work in E-E-S cognition in order to illustrate difficulties that arise when trying to justify a minimal or modified conception of representation through application of ideas like simplicity, unification, and indispensability.

Speaker's Name:Richard Shiffrin
First Author's Name:Andrew Cohen
First Author's Affiliation:Indiana University
Second Author's Name:Adam Sandborn
Second Author's Affiliation:Indiana University
Third Author's Name:Richard Shiffrin
Third Author's Affiliation:Indiana University
Title:Model Selection with Little Data: Analyze by Groups or Individuals?
Abstract:In selecting a best model it is often better to use individual analysis (fitting models separately to individuals) than group analysis (fitting the combined data): Grouping can distort the form of data, and different individuals might perform the task using different processes and parameters. This situation can change dramatically when there is only a small amount of data per individual, due to the possible introduction of distortions and biases into individual analyses. We show this with a simulation technique in which data are generated from each of two known models, each with parameter variation across simulated individuals. We examine how well the generating model and its competitor each fare in fitting (both sets of) the data, using both individual and group analysis. We examine the accuracy of model selection (the probability that the generating model will be selected by the analysis method.) Trials per condition and individuals per experiment are varied systematically. Four pairs of cognitive models were compared: exponential vs. power models of forgetting; GCM (two versions) vs. prototype models of categorization; and FLMP vs. LIM models of information integration.

Speaker's Name:Sverker Sikström
First Author's Name:Sverker Sikström
First Author's Affiliation:Lund University Cognitive Science
Title:Exploring the High Dimensional Semantic Space in the Brain
Abstract:Processing of words from semantic word classes activates networks of semantic representations in the human brain. Earlier research has investigated this by subtracting brain activity evoked from two semantic word categories chosen prior to the experiment. Here we show that arbitrary semantic representations in the brain can be investigated by utilizing high dimensional semantic spaces, which can be generated from the information of co-occurrence in huge text corpora. This method is applied to data were subjects study words list during EEG recording. We correlate estimates of semantic distance with ERP potentials to study semantic representations in the brain. The results show that a large number of different semantic categories show specific topographical patterns across time. This method has a number advantages including studying of arbitrary word classes, and single concepts. Furthermore, semantic space does not require a non-semantic control condition, nor trial-by-trial matching between the psychological construct and brain measure.

Speaker's Name:Mark Steyvers
First Author's Name:Mark Steyvers
First Author's Affiliation:University of California, Irvine
Second Author's Name:Pernille Hemmer
Second Author's Affiliation:University of California, Irvine
Title:A Bayesian Model for Reconstructive Memory
Abstract:It is well established that prior knowledge influences reconstruction from memory, but the specific interactions of memory and knowledge are unclear. Extending work by Huttenlocher et al. (1991, 2000) we propose a hierarchical Bayesian model of reconstructive memory in which prior knowledge interacts with episodic memory at multiple levels of abstraction. The combination of prior knowledge and noisy memory representations is dependent on familiarity. We present empirical evidence of the hierarchical influences of prior knowledge, showing that the reconstruction of familiar objects is influenced toward the specific prior for that object, while unfamiliar objects are influenced toward the overall category.