Slide Sessions

Slides are presented in seven sessions with eight presentations each two-hour session. The format of the slide sessions follows the SfN format: 12 minutes for the talk plus 3 minutes for questions (a total of 15 minutes per speaker).

Session 1: Emotion and Social Cognition

Saturday, April 17, 3:00 - 5:00 pm, Westmount Ballroom

Session 2: Thinking and Decision Making

Saturday, April 17, 3:00 - 5:00 pm, Outremont Ballroom

Session 3: Language

Sunday, April 18, 10:00 am - 12:00 pm, Outremont Ballroom

Session 4: Long-Term Memory

Monday, April 19, 10:00 am - 12:00 pm, Outremont Ballroom

Session 5: Working Memory and Executive Functions

Monday, April 19, 3:00 - 5:00 pm, Outremont Ballroom

Session 6: Attention

Tuesday, April 20, 10:00 am - 12:00 pm, Outremont Ballroom

Session 7: Perception & Action

Tuesday, April 20, 1:00 - 3:00 pm, Outremont Ballroom

Slide Session 1
Saturday, April 17, 3:00 - 5:00 pm, Westmount et al Ballroom

Emotion and Social Cognition

Chair: Kevin Ochsner, Columbia University

Presentation 1: Opponent coding of emotional facial categories

Joshua M. Susskind¹, Adam K. Anderson¹; ¹University of Toronto

Following Darwin’s early observations (Darwin, 1872), it is now widely thought that facial expressions are recognized in relation to one of six or more independent prototypes, with cross cultural, neuroimaging, and neuropsychological studies supporting these prototypes as the fundamental building blocks of emotional representation. However, Darwin’s original thesis was that emotional expressions are not distinct but rather are derived from opposing pairs for efficient social transmission. Here we provide evidence for such an opponent coding model of facial expression appearance and perception. Employing a computational model to form a multidimensional face space representing contrasting facial actions, we examine both behavioral and neural indices of opponent coding predicted by the model. First we examine model and human emotion judgments and demonstrate that visual-statistical opposites of expressions (i.e., emotional anti-faces) signal opposing emotions. Second we show that contrasting emotional expressions cancel each other’s appearance, nullifying subjective intensity judgments and objective discrimination measures of expression perception. Third we show that comparing physically opposing emotional expressions enhances discrimination. Finally, we examine neural indices of opponent coding and show that adapting to an expression results in opposing afterimages in an identical test stimulus, biasing perception of the center of expression space toward its opposite. These converging sources of evidence suggest that facial expressions are not distinct, but rest upon underlying opposing contrasts encoded via opponent neural representations, similar to that of basic perceptual attributes such as color and motion, serving to enhance detection of deviations from neutral states.

Presentation 2: Stressed men and women process faces differently

Mara Mather¹, Nichole R. Lighthall¹, Lin Nga¹, Marissa A. Gorlick¹; ¹University of Southern California

Under stress, men tend to withdraw socially while women seek social support. In the current study, we examined whether these sex differences in social behavior under stress extend to one of the most basic social transactions: face processing. Participants assigned to the stress group were asked to hold their hand in ice water for three minutes (control participants held their hand in warm water). This stress manipulation increased cortisol levels. During the period of post-stress elevated cortisol, we conducted a 3 T functional magnetic resonance imaging scan of participants while they viewed angry and neutral faces. Participants also completed a separate face localizer scan. We found that fusiform face area (FFA) response to faces was diminished by acute stress in males but increased by stress in females. Furthermore, in stressed males looking at angry faces, the insula showed reduced coordination with the FFA and the amygdala, but the inverse was true in stressed females. The insula plays a key role in mirroring the emotions of others and helps us understand what others feel. Thus, our findings that stress reduces insula involvement in brain networks that process emotional expressions in males but increases insula involvement in those same networks in females have important implications for social interactions and empathy under stress.

Presentation 3: Unconscious Affect and Economics Behavior: fMRI Evidence for Stronger Unconscious than Conscious Affective Influences on Financial Choices

Julie L. Hall¹, Richard Gonzalez¹, Chandra Sripada¹, Oliver C. Schultheiss²; ¹University of Michigan, ²Friedrich-Alexander University

Traditionally, the dominant method of measuring affect in psychological research has been through the use of explicit, self-report measures. More recently, implicit, unconscious measures have been developed to assess emotions that respondents may not be consciously aware of or able to report. The current study provides evidence for stronger unconscious than conscious affective influences on financial choices. Using fMRI, 24 participants viewed happy, angry, and neutral affective primes presented under subliminal and supraliminal conditions followed by an investment task where they had to decide between risky, high-payoff stocks and safe, low-payoff bonds. Our results indicate that both subliminal and supraliminal presentations of affective primes influence financial investment decisions and anticipatory neural activation in the nucleus accumbens. As predicted, participants showed greater nucleus accumbens activation and made more risky investment decisions after happy versus neutral face primes, an effect that was significantly stronger for subliminal versus supraliminal primes. Furthermore, implicit measures of positive affect were associated with more nucleus accumbens activation and risky investment decisions and implicit measures of negative affect were associated with more safe bond choices; on the other hand, explicit measures of positive and negative affect had no effect on financial choices or anticipatory neural activation. In conclusion, our results demonstrate that affect influences financial investment decisions and neural markers of anticipatory arousal, particularly at an unconscious level.

Presentation 4: Prefrontal-subcortical pathways mediating successful regulation of craving in cigarette smokers

Hedy Kober^1,2, Peter Mende-Siedlecki³, Ethan Kross⁴, Jochen Weber², Walter Mischel², Carl Hart^2,5, Kevin Ochsner²; ¹Yale University, Department of Psychiatry, New Haven, CT, ²Columbia University, Department of Psychology, New York, NY, ³Princeton University, Department of Psychology, Princeton NJ, ⁴University of Michigan, Department of Psychology, Ann Arbor, MI, ⁵New York State Psychiatric Institute, New York, NY, ⁶

Substance use disorders are chronic relapsing conditions that represent a societal problem with staggering social costs. Drug craving has long been considered a primary contributor to drug use, and the inability to regulate it is thought to be at the root of compulsive drug-taking behavior. Although the neural systems underlying drug craving have been described, this work has yet to make contact with research on the neural systems supporting the ability to regulate emotion. We bridged this knowledge gap to examine neural activity in the most prevalent substance abusing population in the US – cigarette smokers – as they used cognitive strategies to regulate craving for cigarettes and food. We previously reported that the cognitive down-regulation of craving was associated with activity in regions previously associated with regulating negative emotion including dorsomedial, dorsolateral, and ventrolateral prefrontal cortex. This was accompanied by decreased activity in regions previously associated with craving, including the ventral striatum (VS), subgenual cingulate, amygdala, and ventral tegmental area. Here we report that decreases in craving correlated with decreases in VS activity and increases in DLPFC activity. Importantly, VS activity fully mediated the relationship between dlPFC and reported craving. These results provide insight into the mechanisms that enable cognitive strategies to effectively regulate craving, suggesting that it involves neural dynamics parallel to those involved in regulating other emotions. In so doing, this study provides a methodological tool and conceptual foundation for studying this ability across substance abusing populations and developing more effective treatments for substance use disorders.

Presentation 5: Degree of familiarity with situations and people influences the functional overlap between theory of mind and autobiographical memory.

Jennifer S. Rabin¹, R. Shayna Rosenbaum^1,2; ¹York University, ²Rotman Research Institute

Neuroimaging studies find that overlapping and unique brain regions underlie autobiographical memory (AM) and theory of mind (ToM). Investigations of people who lack AM but have intact ToM suggest a dissociation between these abilities; however, this is based on standard ToM tasks that may not induce simulation processes due to their impersonal nature. Greater functional overlap is expected between AM and ToM in known situations involving familiar people that may facilitate simulation of another’s mental state. Here, we aimed to clarify the functional relationship between AM and ToM by varying the degree of personal experience with the protagonist in the ToM condition. Participants were scanned as they remembered past experiences in response to personal photos (AM condition) and imagined others’ experiences in response to photos of personally familiar (friends and relatives, i.e., personal ToM condition; pToM) and unfamiliar people (i.e., impersonal ToM condition; iToM). Analyses revealed common activity within bilateral medial-temporal lobes, medial prefrontal cortex (PFC), posterior cingulate/retrosplenial cortex, left ventrolateral PFC, and left middle temporal gyrus in all three conditions. Functional overlap was greatest during AM and pToM. Furthermore, AM and pToM elicited greater activation of medial PFC and posterior regions relative to iToM, whereas pToM and iToM elicited greater activation of lateral temporal, left ventrolateral PFC, and temporal-parietal regions relative to AM. These data suggest that ToM tasks involving familiar people may induce simulation processes that draw on the same neural network as AM and that disruption to this network may lead to both AM and ToM compromise.

Presentation 6: Ingroup bias in neural empathy predicts ingroup favoritism in costly helping

Grit Hein¹, Giorgia Silani¹, Kerstin Preuschoff¹, Daniel C. Batson², Tania Singer¹; ¹Laboratory for Social and Neural Systems Research, University of Zürich, ²Department of Psychology, University of Kansas

Often, people are more willing to engage in prosocial behavior towards “their people” than members of a different group, causing serious social conflicts. Here, we provide a neurobiological mechanism for ingroup favoritism in helping behavior. We measured the brain responses of soccer fans while they watched either a fan of their own favorite team (ingroup member) or a fan of a rival team (outgroup member) suffering pain. In a second session, scanned participants were given the opportunity to decide, among other options, whether to help the ingroup and outgroup member by opting to endure half of their pain themselves. Thus, helping was costly because it entailed receiving a moderately painful shock. Participants’ empathic brain responses in left anterior insula were greater when they saw an ingroup, as compared to an outgroup, member suffering. The extent of this ingroup bias in empathic brain responses predicted participants’ degree of ingroup favoritism in subsequent helping. Moreover, participants with a negative impression of the outgroup member showed nucleus accumbens (NAcc) activation instead of empathic responses in anterior insula, when they saw an outgroup member in pain. The strength of the signal in NAcc -previously linked to Schadenfreude- was correlated with reduced helping of the outgroup member. These findings stress the importance of cultivating empathy for outgroup members in order to counteract ingroup favoritism and to promote prosocial behavior towards all human beings.

Presentation 7 (GSP Winner): Neurochemical modulation of human prosocial behavior and moral judgment

Molly Crockett¹, Luke Clark¹, Marc Hauser², Trevor Robbins¹; ¹University of Cambridge, ²Harvard University

Studies on the neural basis of human morality suggest that our ability to “know” right from wrong arises from mechanisms distinct from those that enable us to “feel” right from wrong and guide our social behavior appropriately. In the current study, we differentially manipulated moral judgment and prosocial behavior in healthy volunteers by altering neurotransmitter function with citalopram, which boosts serotonin function, and atomoxetine, which augments prefrontal noradrenaline and dopamine function. We administered citalopram, atomoxetine, and placebo on three separate occasions to 30 healthy volunteers in a double-blind, placebo-controlled, fully counterbalanced crossover design. Ninety minutes following drug administration, volunteers played several one-shot ultimatum games, in which they decided whether to accept or reject fair or unfair monetary offers from another player, and judged a standard set of moral dilemmas that included emotion-provoking ‘personal’ moral scenarios, less emotional ‘impersonal’ moral scenarios, and non-moral control scenarios. Citalopram increased prosocial behaviour in the ultimatum game, relative to both atomoxetine and placebo, but atomoxetine had no effect on prosocial behaviour. Meanwhile, atomoxetine and citalopram had distinct effects on moral judgment: relative to placebo, atomoxetine reduced the likelihood of judging both personal and impersonal harms as permissible, while citalopram reduced only the likelihood of judging personal moral harms as permissible. Finally, across individuals the tendency to reject unfair offers in the ultimatum game was significantly correlated with the tendency to judge personal, but not impersonal harms as permissible. These results support the notion of distinct neural mechanisms for moral judgment and prosocial behaviour.

Presentation 8: The insula in political judgment

Bradley Thomas¹, Daniel Tranel¹, Michael Alvarez², Kyle Mattes¹, Ralph Adolphs², Michael Spezio²; ¹University of Iowa, ²Caltech

Social judgments made solely on the basis of a candidate’s visual appearance correlate with real-world election outcomes (Todorov et al., 2005). We previously showed that this effect is associated with regional activation of the insula during such judgments, and in particular that the insula is more activated when viewing people who lost real elections, especially when participants are primed to process threat-related information (Spezio et al., 2008). Here we investigated the causal role of the insula with a lesion study. Using the same stimuli and tasks as in our prior study (Spezio et al., 2008), we tested 12 individuals with focal lesions to the insula. The insula-lesioned group showed normal age judgments of the faces that were associated with real-world election wins, as is the case in healthy individuals (t=2.6; p<0.02). They also showed a trend association with real-world elections when asked to carry out a simulated vote (t=1.5, p=0.08). However, whereas healthy individuals show a robust association of real-world election outcomes with ratings of competence (positively) and threat (negatively), the insula-lesioned group showed neither effect (t=-0.67, p=0.3 for threat; t=0.54, p=0.3 for competence). The findings suggest that the insula is required for threat-related cues from a candidate’s appearance to influence election outcomes.

Slide Session 2
Saturday, April 17, 3:00 - 5:00 pm, Outremont et al Ballroom

Thinking

Chair: Kalina Christoff, University of British Columbia

Presentation 1 (GSP Winner): Is Morality Unified? The Neural Correlates of Different Kinds of Moral Judgments

Carolyn Parkinson¹, Philipp E. Koralus², Walter Sinnott-Armstrong¹, Angela Mendelovici², Victoria McGeer², Thalia Wheatley¹; ¹Dartmouth College, ²Princeton University

Researchers have recently begun to investigate moral judgment using neuroscientific techniques. By and large, these studies have been designed assuming that morality is sufficiently unified to be studied as a single kind of judgment. However, scholars in other fields (e.g., philosophy, anthropology) have posited that there are multiple distinct domains of morality. The current study used functional imaging to examine the neural correlates of the judgment of 3 categories of moral transgressions: dishonest, disgusting and harmful acts. 40 naïve, right-handed adults were scanned in a 3.0-Tesla Phillips scanner. While in the scanner, subjects read 14 morally ambiguous scenarios from each of 3 moral categories (disgust, dishonesty, harm), as well as 14 neutral scenarios. After reading each scenario, subjects used a button press to indicate whether the act described was morally ‘wrong’ or ‘not wrong.’ Judgments of moral transgressions that evoked disgust elicited increased activity in the insula, amygdala, and cingulate cortices. The judgment of dishonest acts was correlated with increased bilateral activity in the temporoparietal junction, while judgment of physically harmful acts was correlated with increased activity in the inferior parietal lobule, insula, precuneus and posterior cingulate. These results contrast with those of past studies that have lumped so-called ‘moral judgments’ together, and suggest that judgments of moral scenarios involving different kinds of transgressions draw on disparate neural and cognitive mechanisms. Thus, future studies should differentiate between subsets of moral judgments, and identify factors that unify and distinguish them from one another.

Presentation 2: Mechanisms Underlying Strategic Variability in Economic Decision Making

Vinod Venkatraman¹, John W. Payne¹, Michael W.L. Chee², Scott A. Huettel¹; ¹Duke University, ²Duke-NUS Graduate Medical School, Singapore

Adaptive decision making involves the use of multiple strategies that vary with decision context, both within and across individuals. In a series of behavioral, eye-tracking and fMRI experiments using an economic decision-making task, we sought to understand the mechanisms underlying strategic variability in risky choice. In each trial, participants were provided with a three- or five-outcome mixed gamble consisting of at least one monetary gain, one monetary loss and one intermediate outcome that was typically $0 or slightly negative. They could make three types of choices: gain-maximizing (increasing the magnitude of largest gain), loss-minimizing (decreasing the magnitude of worst loss) or probability-maximizing (increasing overall probability of winning by modifying valence of intermediate outcome). Across multiple studies, we demonstrated a systematic bias towards probability-maximizing choices. Individual differences in the extent of this bias correlated positively with a satisficing trait, negatively with a sadness trait, and could be predicted using patterns by which information about the gambles was acquired (indexed using eye positions). We also show that strategic preferences can be modified using subtle variations in the decision context, as well as following 24 hours of total sleep deprivation. Finally, using fMRI, we demonstrate that distinct neural systems were associated with choices and with strategic preferences across individuals. Moreover, changes in these neural systems predicted an increased bias toward gain-maximizing choices following sleep deprivation. Our findings highlight individual variability in strategic preferences as a key direction through which neuroscience data can influence models of risky choice.

Presentation 3: Circuits for overcoming decision inertia in the human brain

Stephen Fleming¹, Charlotte Thomas², Raymond Dolan¹; ¹University College London, ²University of Bristol

Humans often accept the status quo, or default, when faced with conflicting choice alternatives. However, it is unclear how neural pathways linking cognition and action contribute to such decision inertia. To examine this question, we used a visual detection task where decision difficulty and the default response option were systematically and independently manipulated within a factorial design. We show that subjects tend to accept a default when making difficult, but not easy, decisions. This bias was suboptimal in that more errors were made when the default was accepted. Using functional magnetic resonance imaging (fMRI) we show that an interaction between decision conflict and default acceptance was uniquely expressed in enhanced activity in bilateral regions that encompass the subthalamic nuclei (STN). These findings suggest a specific role for STN activity in overcoming conflict-induced decision inertia. Consequently, to reveal the mechanism driving this effect we employed an effective connectivity analysis to show that right inferior frontal cortex is both sensitive to conflict and modulates the STN, contributing to decision inertia. Importantly, a driving influence of right inferior frontal cortex on the STN was increased during rejection of the default, supporting a role for this pathway in overcoming a status quo bias. We conclude that overcoming inertia during difficult decisions invokes specific neural dynamics within prefrontal-basal ganglia circuitry, suggesting commonalities between the neural mediation of the status quo bias in healthy individuals and decision inertia induced by pathologies of movement.

Presentation 4: Brain Stimulation to the Parietal and Dorsolateral Prefrontal Cortices Reveals Double Dissociation between Learning and Automaticity

Roi Cohen Kadosh¹, Sonja Soskic², Teresa Iuculano², Ryota Kanai², Vincent Walsh²; ¹University of Oxford, ²University College London

How do we learn to encode and represent numerical information? What are the necessary brain mechanisms underlying such abilities? Cognitive, neuropsychological, and anthropological studies show that intact and proficient numerical representation is characterised by: 1) automatic retrieval of numerical quantity, and 2) interactions between number and space, such as for example accurate mapping of numbers on a physical line. We examined here how non-invasive brain stimulation (transcranial direct current stimulation (tDCS)) of the right dorsolateral prefrontal (DLPFC) and the right posterior parietal (PPC) cortices over 6 days of numerical learning affects: 1) the efficiency of learning of a new numerical system, 2) the development of its automaticity, and 3) the interaction between newly learnt numbers and space. The current results revealed a double dissociation between the DLPFC and the PPC. The DLPFC was crucial for numerical learning, while the PPC was necessary for the acquisition of automatic numerical quantity and its interaction with space. These effects were specific to the learning of a new numerical system as brain stimulation did not affect the automaticity and the mapping of numbers into space of everyday digits. These results challenge the idea that automaticity is a result of effective skill acquisition. Furthermore, they question the necessity of the DLPFC in symbolic numerical representation, and have implications for models of typical and atypical numerical development.

Presentation 5: Critical connectivity: a neurobiological threshold effect for creative ability

Rachael Grazioplene¹, Robert S. Chavez¹, Alison Marshall¹, Ranee A. Flores¹, Rex E. Jung¹; ¹The Mind Research Network

The threshold theory of creativity posits that intelligence predicts performance on divergent thinking tasks, but only up to an IQ of about 120 (Runco and Albert, 1986). However, a meta-analysis casts doubt on the validity of the threshold effect (Kim, 2005). These contradictory behavioral findings might be refined if we understand the relationship between brain architecture and creativity measures at different IQ levels. Therefore, we hypothesized differential white matter-behavior relationships in subjects below an IQ of 120 compared to those above 120. A group of healthy normal volunteers (n=100, ages 18-29) underwent cognitive testing and structural MRI scanning at 3 Tesla. Based on the mean Full-Scale Intelligence Quotient (FSIQ) of 119, the sample was split into 2 groups for analysis (Lower Group<120, n=47; Higher Group>=120, n=53). Behaviorally, FSIQ conformed to a linear relationship with divergent thinking scores in the Lower FSIQ Group (r=0.32, p<0.05), but not in the Higher FSIQ Group. Voxel-wise cross-subject statistical analyses showed a strong inverse correlation between frontal white matter fractional anisotropy (FA) and composite divergent thinking scores in the Higher FSIQ Group (clusters significant at p=0.01-0.001); no such relationship approached significance in the Lower Group. This finding suggests that lower white matter coherence in frontal thalamo-cortical white matter may confer cognitive advantages for divergent thinking abilities, but only in conjunction with intelligence levels in the superior range and higher (FSIQ >120).

Presentation 6: Spontaneous and deliberate modes of thought during the creative process

Melissa Ellamil¹, Charles Dobson², Mark Beeman³, Kalina Christoff¹; ¹University of British Columbia, ²Emily Carr University of Art and Design, ³Northwestern University

Until recently, neuroscientific investigations of human thought equated it with only the deliberate, goal-directed mental processes that occur during problem solving, while spontaneous mental processes (e.g., mind wandering) occurring without direct control have been overlooked. The creative process appears to combine both modes of thinking: the generation of novel ideas is facilitated by low cognitive control and defocused attention that characterize spontaneous thought, while the evaluation of those ideas is aided by high cognitive control and focused attention that characterize deliberate thought. Deliberate thought has been linked to the brain’s executive network, while spontaneous thought seems to activate the default network. However, both modes of thought have so far not been examined within the same study, and their neural correlates remain unclear. The study used a creativity task during functional magnetic resonance imaging (fMRI) to examine neural recruitment during spontaneous and deliberate modes of thought in the context of creative production. Using an fMRI scanner-compatible drawing tablet, students from a local arts college alternated between generating ideas for a book cover, evaluating the ideas generated, and tracing over lines that appeared on the screen (baseline). Generation was associated with activations in the default (parahippocampus) and executive (ventrolateral prefrontal cortex [PFC]) networks. Evaluation was also associated with activations in the default (medial PFC, posterior cingulate cortex/precuneus) and executive (dorso/rostrolateral PFC, dorsal anterior cingulate cortex) networks. Thus, in the context of creative production, spontaneous and deliberate modes of thought demonstrate parallel recruitment of two neural networks thought to work in opposition.

Presentation 7: White matter pathways associated with verbal episodic and working memory, measured using tract-based spatial statistics in normal aging

Rebecca Charlton¹, Thomas Barrick¹, Hugh Markus¹, Robin Morris²; ¹St George's University of London, ²Institute of Psychiatry, King's College London

Diffusion tensor imaging (DTI) is sensitive to loss of white matter integrity in normal aging. Using whole brain measurements and large regions of interest, DTI has been shown to correlate with both verbal working (WkM) and episodic memory (EM). The impact of changes in white matter integrity on the networks that support these functions remains unclear. It has been suggested that these memory abilities utilise common frontal grey matter regions, but also unique grey matter in the parietal (WkM) and temporal (EM) lobes. To explore this further tract based spatial statistics (TBSS), which allows 3D investigation of white matter tracts, was performed on 98 middle-aged and older adults. White matter tracts were identified that were associated with variations in WkM and EM in the frontal lobes bilaterally. In addition, verbal WkM was associated with white matter pathways in the left parietal lobe, whereas verbal EM was associated with bilateral temporal pathways. The results support the notion that damage to certain white matter pathways may have widespread effects on different cognitive abilities by causing disruption to multiple distributed networks. In addition, certain abilities may be more affected by damage to specific tracts, because tract disruption has specific effects on a unique network.

Presentation 8: Levels of difficulty in autobiographical and visuospatial planning differentially modulate core, dorsal attention and control network activity

R. Nathan Spreng¹, Adrian W. Gilmore¹, Daniel L. Schacter¹; ¹Harvard University

Planning is an adaptive function that guides complex everyday behavior. We recently reported a novel autobiographical planning (AP) task where participants were given a goal state and integrated three component steps into a personal plan to fulfill that goal. We found that AP engaged the core network, including frontal and posterior midline structures, medial and lateral temporal lobes and posterior inferior parietal cortex. Visuospatial planning, as measured by the Tower of London (TOL) task, engaged the dorsal attention network, including dorsolateral prefrontal cortex, frontal eye fields, intraparietal sulcus, and ventral occipital cortex. Relative to a counting task, both planning tasks engaged the frontoparietal control network (FPCN), including rostrolateral prefrontal cortex, anterior cingulate, and anterior inferior parietal cortex. To assess the impact of task demands on this pattern of activity, we examined changes in BOLD signal, using fMRI in 25 participants, associated with parametrically modulating the difficulty of both autobiographical and TOL planning. Compared with easy trials, difficult AP trials required the integration of more steps and difficult TOL trials required more moves to reach a solution. Easier AP elicited greater activity in the core network than difficult AP. TOL difficulty modulated the magnitude of dorsal attention network activity, consistent with previous reports. The FPCN was engaged in both planning tasks. However, easier AP engaged the FPCN less than difficult AP. TOL task difficulty did not modulate FPCN activity. These findings expand our understanding of the functional interaction between the FPCN and the core and dorsal networks related to task demands.

Slide Session 3
Sunday, April 18, 10:00 am - 12:00 pm, Outremont et al Ballroom

Language

Chair: Tamara Swaab, University of California, Davis

Presentation 1 (GSP Winner): Temporal ERP evidence for an early dissociation between semantic and lexical representations in function of task demands

Kristof Strijkers^1,2, Phillip Holcomb³, Albert Costa²; ¹Universitat Pompeu Fabra, Spain, ²Universitat de Barcelona, Spain, ³Tufts University

Most models on object processing and naming posit that activation automatically cascades from semantic processing to lexico-semantic and, according to some, word form representations. Given this continuous transmission of information, top-down processing in function of task demands and context should at some point in time gate the flow of activation towards those representations relevant for response execution. However, very little is known about how and especially when top-down task demands affect bottom-up object processing. Based on recent findings showing early lexical ERP modulations (~ 175 – 200 ms) in overt naming tasks (Strijkers et al., 2009; Costa et al., in press), we compared the time-course of processing during overt picture naming (experiment 1) with that of go/no-go semantic categorization (experiment 2), assuming that only naming requires fast lexical activation. In order to obtain a reliable measure of access to the object’s name, lexical frequency was manipulated. During overt naming, low frequency ERPs diverged from high frequency ERPs very early on (~ 175 ms), replicating recent findings. Importantly, during semantic categorization of the same objects, low frequency and high frequency ERPs did not yield any early modulations. These results offer strong additional evidence that the brain rapidly accesses the lexical system after perceiving a picture; however, the fast lexical engagement only occurs when actual naming is required. This is the first direct on-line demonstration for an early functional dissociation between the semantic and lexical systems, with top-down task demands shaping bottom-up activation earlier as expected by most theories.

Presentation 2: The time course of orthography and phonology during reading aloud in a second language: Evidence from an ERP study

Kalinka Timmer^1,2, Niels O. Schiller^1,2; ¹Leiden University, ²Leiden Institute for Brain and Cognition

This study confirmed that the Masked Onset Priming Effect (MOPE), which is often found in primary language usage, can also be observed in a second language. Bilingual Dutch-English participants read aloud English words starting with phonologically ambiguous letters (e.g. /c/, carpet or cigar) or with phonemes that have multiple orthographies (e.g. /f/, phase or fellow) in a masked priming experiment to separate the contribution of orthographical and phonological activation in reading aloud. Orthographical priming was found in early time windows of event-related potentials (ERPs), even if it did not affect the overall reaction times. Phonological priming, on the other hand, shortened the response time and was weakly indicated in later time windows of ERPs. Both the orthographic and phonological priming effects were present first for unambiguous onset words and later for ambiguous onset words, where the onset has multiple phonemic possibilities. These results show that non-native speakers, even if they speak with an accent (as they are accustomed to a different phonetic system) and even if they are less familiar to the orthography of their second language, behave in a similar way as native speakers.

Presentation 3: The emergence of sensorimotor representations in the developing brain during language processing

Karin Harman James¹, Shelley Swain¹, Harinder Kaur¹; ¹Indiana University

In adults, verb perception recruits motor systems – we have recently shown that motor systems are also active when preschool children perceive verbs. Although the assumption is that our history of actions leads to this activation, until now, there has been no empirical evidence to support this claim. The present research uncovers experiences that are required for this motor recruitment to occur. Twelve children, aged 5-7 years, learned novel labels (‘made-up’ verbs) for actions performed with novel objects. Half of these labels were learned during ‘active’ interaction – the children manipulated the objects while the verb label was given, the other labels were learned ‘passively’ - while the children watched the experimenter demonstrate the actions. Subsequent to this experience, children heard all the novel verbs while BOLD activation was measured using fMRI. Results demonstrated that motor systems were recruited during ‘verb’ perception only after the participants learned the labels during active exploration. In contrast, no motor activation emerged when labels were learned by passively watching the experimenter during learning. These results demonstrate that verb perception recruits motor systems due to active involvement with objects. Language learning through active interaction with the environment results in representations that incorporate both motor and sensory systems in the developing brain.

Presentation 4: Left occipital cortex of congenitally blind adults responds to grammatical structure

Marina Bedny^1,2, David Feder¹, Evelina Fedorenko¹, Elizabeth Hawkins¹, Nancy Kanwisher¹, Alvaro Pascual-Leone², Saxe Rebecca¹; ¹Massachusetts Institute of Technology, ²Berenson-Allen Center for Noninvasive Brain Stimulation, Beth Israel Deaconess Medical Center, Harvard Medical School

In typically developing adults, language processing is supported by left prefrontal and lateral temporal cortices. In addition, to these brain regions, congenitally blind individuals activate parts of the left occipital cortex, including the primary visual cortices, during language tasks. The role of left occipital cortex in language processing is unclear. We studied the function of the left occipital cortices in three experiments with congenitally blind adults and sighted controls. In Experiment 1 congenitally blind and sighted participants listened to stories and answered true/false questions about them. In a control task, participants made a match/non-match judgment with backwards speech sounds. We found that the left occipital cortex responded more to sentences than backwards speech, despite the greater difficulty of the backwards speech task as measured by reaction time and accuracy. Experiment 2 showed that that left occipital cortex is not involved in retrieving word meanings: left occipital activity was not observed during a single-word comprehension task. Experiment 3 further examined the role of the left occipital cortex computing grammatical structure by comparing sentences, word lists, Jabberwocky sentences, lists of non-words and backwards speech. We found that left occipital cortex responded more to sentences than all other conditions and more to Jabberwocky sentences than to word lists, non-word lists and backwards speech. Word-lists, non-word lists and backwards speech conditions were not different from each other. We conclude that the left occipital cortices of congenitally blind adults are involved in constructing syntactic, sentence-level representations.

Presentation 5: How the brain predicts forthcoming words during sentence listening

Jeremy Skipper¹, Jason Zevin¹; ¹Sackler Institute for Developmental Psychobiology, Weill Medical College of Cornell University

In natural settings, listeners use sentence context to predict forthcoming discourse content. We propose that hypotheses about upcoming words are specified as motor commands that would be used to produce those words and that efference copy from the motor system predicts the occurrence of actual words. We tested this model using fMRI while participants listened to sentences that varied in final word predictability. High and low-predictability sentences were composed of a sentence frame, a filled pause of variable duration, and a final word. If listeners use sentence context to predict words, we expected "repetition suppression" of activity during the final word relative to the pause for high-predictability sentences in language comprehension areas. Indeed, anterior superior temporal cortex and the pars triangularis showed repetition suppression. Additionally, the spatial pattern of activity during the high-predictability pause was most correlated with the pattern of activity for the low-predictability final word in anterior superior temporal cortex only. Finally, if words were activated through efference copy from the motor system then activity during the high-predictability pause should occur in speech production areas. Indeed, inferior parietal, insula, and premotor cortex showed greater activity during the high compared to the low-predictability pause. Results suggest that the brain uses knowledge and expectations about sentence context to activate forthcoming words through efference copy from areas of cortex involved in speech production. We suggest that this predictive mechanism constrains interpretation and allows the brain to expend less time processing incoming sensory information, allowing resources to be allocated elsewhere.

Presentation 6: The neural basis of definite reference in discourse – an fMRI study

Amit Almor¹, Timothy Boiteau¹; ¹University of South Carolina

Anaphoric reference is an important part of coherent discourse but very little is known about the brain basis of processing discourse anaphors. Much research in Linguistics and Psycholinguistics has indicated that anaphoric references to salient antecedents tend to employ more general expressions than the antecedent, often a pronominal form, but sometimes also a general category term. One intriguing finding about category anaphor processing is the Inverse Typicality Effect, which is the slower reading of a category anaphor (e.g., the bird) when referring to a salient antecedent, when that antecedent is a typical member of the category (e.g., robin) rather than an atypical one (e.g., chicken). This effect was previously argued to reflect increased working memory load caused by interference between the semantic representations of the category anaphor and of the salient typical member antecedent (Almor, 1999). Here we used fMRI to test this claim by examining the brain regions that are differentially activated by discourses with category anaphors with salient typical antecedents, salient atypical antecedents, and non salient typical antecedents. We found that category anaphors with salient antecedents elicited greater activation with typical antecedents than with atypical antecedents in IFG bilaterally and in LH Fusiform Gyrus. The IFG activations are consistent with a WM involvement in the Inverse Typicality Effect while the Fusiform activations suggest that this effect also involves activations of lexical representation. The lack of activation differences between the conditions in other temporal regions suggests that the representations involved in this effect are lexical-semantic rather than conceptual-semantic.

Presentation 7: A Diffusion Tensor Imaging Investigation of Individual Differences in White Matter Microstructure as a function of Reading Skill and Working Memory Capacity

Chantel Prat¹, Sarah Schipul¹, Timothy Keller¹, Marcel Just¹; ¹Carnegie Mellon University

This study used diffusion tensor imaging (DTI) to investigate individual differences in the organization of white matter microstructure as a function of two behavioral measures of language ability. Associations of fractional anisotropy (FA) with indices of reading skill (Nelson-Denny Reading Test) and working memory capacity (Reading Span) were examined in a voxel-based multiple regression analysis of 73 healthy adults (40 males, aged 18-31). Both reading skill and working memory capacity (which were not correlated with one another) were positively correlated with FA, suggesting increased microstructural organization in higher-skilled and higher-capacity individuals. Higher reading skill was associated with increased FA in left hemisphere (LH) regions of superior longitudinal and inferior fronto-occipital fasciculi (connecting Broca’s and Wernicke’s areas) and in the anterior thalamic radiation (connecting thalamus to medial prefrontal regions). Higher working memory capacity was associated with increased FA in the forceps minor (connecting LH and RH frontal regions) and the cingulum (connecting the frontal, parietal, and temporal lobes). Increased FA in the LH of skilled readers may reflect increased myelination arising from more reading practice. Previous research has demonstrated such an increase in LH FA with intensive reading exposure (Keller & Just, in press). Increased FA in high-capacity individuals may speed information transfer between the frontal lobes, which are important for working memory and cognitive control, and the rest of the cortex, facilitating higher-level cognitive processes. These results show that systematic structural connectivity variations can be observed in healthy adults as a function of individual differences in cognition.

Presentation 8: The relationship between story comprehension and social comprehension: Quantitative meta-analyses using Activation Likelihood Estimation

Raymond Mar¹; ¹York University

While initial neuroimaging investigations of language focused on lower-level word and syntactical representations, there has been a growing interest in investigating language at the level of discourse, particularly with respect to stories. One major question that has emerged is the relationship between story comprehension and social comprehension, with some hypothesizing that mental-inference processes are engaged while readers read about story characters and their interactions. Although some recent reviews have noted that the network of brain regions implicated during story comprehension resembles the social cognitive network, no review has directly compared the meta-analytic results for these two processes. Two quantitative meta-analyses were undertaken, employing the activation likelihood estimation (ALE) approach, for fMRI and PET studies of both narrative comprehension and mental inference. A total of 18 narrative comprehension articles were entered into the analysis, representing 22 samples, 293 participants, and 203 foci. For mental inference, 54 articles provided 55 samples, 712 participants, and 440 foci. Mental inference studies were further divided into those that employed stories as stimuli, and those that did not, with separate ALE meta-analyses performed for each group. Identifying the core social cognitive network was achieved by examining the overlap of these two groups, implicating the mPFC, bilateral TPJ, bilateral STS, posterior cingulate, and left IFG. The results for narrative comprehension overlapped with all of these areas, except for the posterior cingulate. While these results are consistent with the idea that mental inference is employed during narrative comprehension, other possibilities will be discussed.

Slide Session 4
Monday, April 19, 10:00 am - 12:00 pm, Outremont et al Ballroom

Long-Term Memory

Chair: Roberto Cabeza, Duke University

Presentation 1: Functional magnetic resonance imaging (fMRI) evidence for multiple color knowledge representations influenced by context and cognitive style

Nina S. Hsu¹, David J.M. Kraemer¹, Robyn T. Oliver², Margaret L. Schlichting¹, Sharon L. Thompson-Schill¹; ¹University of Pennsylvania, ²Roosevelt University

Numerous neuroimaging studies have supported sensorimotor theories of semantic memory by demonstrating overlap or proximity of brain regions sensitive to perception and knowledge retrieval. However, particularly within the domain of color knowledge, we hypothesize that some of the differences in results across studies may be due to the existence of multiple types of color knowledge representations. Moreover, retrieval of these representations can be influenced by both context and individual differences in cognitive style. In Experiment 1, we provide fMRI evidence for multiple representations of object color knowledge by having subjects perform a verbal task, in which context encouraged subjects to retrieve high- or low-resolution information about the colors of named common objects in a blocked experimental design. In the left fusiform, we found more activity during retrieval of high- versus low-resolution color knowledge. We also assessed preference for verbal or visual cognitive style, finding that brain activity in the left lingual gyrus significantly correlated with preference for a visual cognitive style. To eliminate the possibility that strategy accounts for the results, we conducted Experiment 2, which had a randomized design and quicker trials. We also removed the possibility of differential visual system engagement during viewing of written stimuli by presenting auditory stimuli instead. Experiment 2 replicated many of the findings of Experiment 1. Overall, these findings suggest the existence of multiple types of color knowledge representations that are based on resolution of detail, and that retrieval of these representations can be influenced by context and individual preferences in cognitive style.

Presentation 2: Asymmetrical resting-state functional connectivity of parahippocampal cortex is associated with laterality differences for form-specific vs. form-abstract repetition priming of scenes

W. Dale Stevens^1,2, Randy L. Buckner^1,2,3, Daniel L. Schacter^1,2; ¹Harvard University, ²Athinoula A. Martinos Center for Biomedical Imaging, ³Howard Hughes Medical Institute

Behavioral priming refers to an improvement in performance upon successive encounters with the same or related stimuli, and is typically accompanied by reduced neural activity in specific cortical regions (neural priming). Neural priming reflects category-specificity, e.g., in bilateral scene-preferential parahippocampal cortex (SP-PHC) for repeated scenes. We recently reported functional asymmetry within SP-PHC: form-specific vs. form-abstract priming were lateralized within the right vs. left hemispheres, respectively, paralleling previous findings in other category-preferential visual regions, e.g., fusiform cortex for object-priming. Here, we tested the hypothesis that right/left differences are associated with differential functional connectivity of these regions with large-scale perceptual vs. conceptual networks, respectively. Bilateral SP-PHC regions were individually defined in healthy adults using independent functional localizers. Using resting-state functional connectivity analysis of fMRI data (rs-fcMRI) with a priori seed regions in middle occipital gyrus (MOG) and posterior inferior frontal gyrus (pIFG), we replicated two large-scale networks associated with perceptual- and conceptual-priming, respectively, as previously identified (Wig et al., 2009). We observed relatively increased correlation of right SP-PHC with the MOG/perceptual network, and of left SP-PHC with the pIFG/conceptual network. Whole-brain voxelwise analyses further demonstrated functionally-specific differential rs-fcMRI: Correlations were relatively higher for right SP-PHC with posterior and ventral visual regions and aspects of the dorsal attention network; and higher for left SP-PHC with regions of the default and frontoparietal control networks. Our results provide novel evidence that right/left form-specific vs. form-abstract priming asymmetry may be related to specialized differential functional connectivity of these regions with perceptual vs. conceptual large-scale networks, respectively.

Presentation 3: Emotional Salience Differentially Modulates Repetition Suppression in the Lateral and Medial Ventral Visual Stream

Sarah Zweynert^1,2, Torsten Wüstenberg², Kerstin Krauel³, Constanze Seidenbecher¹, Emrah Düzel^3,4, Björn Schott^1,2,3; ¹Leibniz Institute for Neurobiology, Magdeburg, Germany, ²Charité University Hospital, Berlin, Germany, ³Otto von Guericke University, Magdeburg, Germany, ⁴University College London, London, UK

Repetiton suppression (RS) describes a decreasing neural response to stimuli as a function of their repeated presentation. Evidence from studies using faces as stimuli suggests that emotional salience modulates repetition effects in stimulus-related brain regions. Here we investigated the functional neuroanatomy of repetition effects and their emotional modulation in brain regions of the ventral visual stream using event-related functional MRI. 28 young healthy participants performed a simple visual working memory task in which targets and distracters (visual scenes) were presented repeatedly. Targets and distracters could either be neutral images or scenes with negative emotional (aversive) content. We hypothesized that in stimulus-related brain regions (e.g. Amygdala, parahippocampal place area (PPA), fusiform gyrus (FG)) repetition suppression would be stronger for emotional compared to neutral scenes. Both target pictures and task-irrelevant distracter pictures elicited a reliable repetition suppression response (1st – 3rd presentation) in ventral visual stream structures. Within these regions, we identified two separable patterns. In the amygdala and in the lateral fusiform cortex, emotional scenes showed significantly stronger initial brain responses and enhanced repetition suppression. In contrast, in the PPA, in the medial FG, and in the anterior hippocampus, repetition suppression was equally pronounced for emotional and neutral items. Our results suggest the presence of two distinct patterns of repetition suppression in the ventral visual stream, with a lateral portion showing emotional modulation of repetition effects and a medial portion that shows emotion-independent responses to repeated stimuli.

Presentation 4: Human hippocampal theta oscillations and the formation of episodic memories

Bradley Lega¹, Joshua Jacobs¹, Michael Kahana¹; ¹University of Pennsylvania

Using data from neurosurgical patients undergoing invasive monitoring, investigators have demonstrated the functional importance of gamma oscillations in the cortex and hippocampus. Evidence for the importance of theta oscillations has been limited mostly to cortex. Select studies from the human literature suggest that the human correlate of the theta oscillation may be centered below 4 Hz. We present data from the largest existing database of patients with hippocampal depth electrodes performing free recall tasks. 33 patients with a total of 237 depth electrodes are included in this study. We employ a novel oscillation detection algorithm to identify theta oscillations. We identify an oscillation between 2.5 and 5 Hz and another located between 7.5 and 10 Hz. The lower frequency oscillation bears properties more in line with animal evidence of theta band activity. Using a subsequent memory effect analysis, we show that higher power in this frequency band is correlated with successful encoding of episodic memories. The effect matches the magnitude of the effect in the gamma band which has been previously documented. We note a concomitant negative SME, in which power during encoding is decreased for words that are subsequently recalled. The low theta SME was more specific in the dominant hemisphere and persisted throughout the 2 second encoding period. Low theta oscillatory activity also increases in the time window that precedes vocalization of a recalled word. Our data locate a human hippocampal theta oscillation between 2.5 and 5 Hz, and couple this oscillation to a specific SME.

Presentation 5: Modulation of Stress Hormone Levels at the Time of Memory Reactivation: An Opportunity to Change Previously Acquired Memories

Marie-France Marin^1,2,3, Sonia J. Lupien^1,2,3; ¹Center for Studies on Human Stress, ²Fernand-Seguin Research Center, Louis-H. Lafontaine Hospital, ³Université de Montréal

Glucocorticoids (GCs) are a major class of stress hormones known to modulate different memory processes. In general, high levels of GCs enhance memory consolidation whereas both low and high levels impair retrieval. Other studies show that the process of retrieval serves as a reactivation mechanism whereby the memory that is reactivated during the retrieval process is once again sensitive to modifications by environmental manipulations. Therefore, Study 1 investigated the immediate and long-term effects of a stressor on a reactivated memory. Thirty-two healthy participants (16 men) encoded a movie containing neutral and emotional slides. Two days later, they recalled the movie. Half were then exposed to the Trier Social Stress Test, a validated psychosocial stressor, whereas the others read magazines (controls). Memory was re-assessed after stress and five days later. The stressed group recalled significantly more emotional material after stress compared to controls. Moreover, this enhanced emotional memory trace was maintained five days later. Study 2 investigated whether pharmacologically lowering GC levels at the time of reactivation would impact the memory trace in a temporary or a long-lasting manner. Twenty-two healthy men encoded the movie. Three days later, they were assigned to a metyrapone (an inhibitor of GC synthesis) or a placebo condition. Memory was assessed after drug administration and four days later. At both time points, the metyrapone group recalled less emotional material compared to controls. These experiments suggest that variations in GC levels at the time of reactivation can modulate the strength of an emotional memory trace.

Presentation 6 (GSP Winner): Changing the course of memory consolidation through reactivation during sleep

John Rudoy¹, Joel Voss^1,2, Carmen Westerberg¹, Ken Paller¹; ¹Northwestern University, ²University of Illinois at Urbana-Champaign

Initially fragile memories can gain stability via consolidation, but the manner in which sleep contributes to this process is unresolved. Reinstating a learning context during slow-wave sleep enhances retrieval of spatial information learned in that context (Rasch et al., 2007), but it remains unclear whether exposure during sleep to cues associated with newly learned information can selectively enhance the storage of individual memories. In our study, subjects learned to associate each of 50 unique object images with a location on a computer screen prior to a nap. During learning, each object was paired with a characteristic sound cue delivered over a speaker (e.g., cat/meow, kettle/whistle). During non-REM sleep in a subsequent nap, the sound cues for 25 of the objects were presented. After waking, individuals viewed all 50 objects and attempted to position each one in its original location. Accuracy was greater for objects cued during the nap than for those not cued. Additionally, EEG responses to sound cues were sorted into two conditions via a median split on the difference between pre- and post-nap accuracy. Average EEG amplitudes measured over the interval from 600-1000 ms after sound onset were greater for cues associated with less forgetting. Thus, the degree of recall improvement or decline was apparently influenced by sound-induced memory processing during sleep, as indexed by brain potentials. These results show that memory processing during sleep can be highly specific, and that auditory reminders during sleep can be used to target the reactivation and strengthening of individual memories.

Presentation 7: Role of emotion in sleep-dependent modulation of forgetting

Rebecca Spencer^1,2, Sam Bromfield¹, Bengi Baran¹; ¹Department of Psychology, University of Massachusetts, Amherst, ²Neuroscience & Behavior Program, University of Massachusetts, Amherst

Sleep enhances memory consolidation. However, a long-standing proposal (Crick & Mitchinson, Nature, 1983), suggests that sleep also increases forgetting of conflicting information. Similarly, a more recent hypothesis, the Synaptic Homeostasis Hypothesis (e.g., Gilestro, Tononi, & Cirelli, Science, 2009), implies that weaker memories from the day are forgotten over sleep through global synaptic downscaling. We examined off-line changes in forgetting using the retrieval induced forgetting paradigm. Participants studied 240 word pairs formed from 40 cues, each of which was affiliated with 6 targets. Following the study phase, participants practiced retrieval of a subset of pairs, consisting of 20 cues and 3 of their affiliated targets. Following a break either with or without sleep, recall of all 240 pairs was tested. Consistent with previous reports, learning of the practiced pairs was greater following a break with sleep relative to an equivalent interval spent awake. Counter to previous hypotheses, competitive forgetting was greater over the interval spent awake relative to the interval containing sleep. Moreover, forgetting was modulated by REM sleep while memory consolidation of practiced pairs was associated with slow wave sleep. In a second study, we found that sleep-dependent consolidation of practiced pairs was similar for emotionally positive, negative, and neutral stimuli, however forgetting was uniquely modulated by the emotional valence. Together, these results suggest that sleep may play a dual-role in memory processing, reducing forgetting of weak memories during REM while consolidating stronger memories during slow wave sleep.

Presentation 8: Multivariate spatio-temporal partial least squares analysis of age-related changes in brain activity during context memory encoding and retrieval

David Maillet¹, Maria Natasha Rajah¹; ¹McGill University

Compared to healthy young adults, healthy older adults exhibit intact item memory, but impaired memory for spatial and temporal context/source. Neuroimaging studies suggest that age-related functional changes in pre-frontal cortex (PFC) and medial temporal lobes may underlie this impairment. In the present study, we examined brain activity during the encoding and retrieval of item, spatial and temporal contextual memory tasks using face stimuli, in both the young (n=25) and the elderly (n=26). Our goal was to identify the similarities and differences in brain activation during recognition and context tasks between young and elderly, using non-rotated multivariate spatio-temporal partial least squares analysis (PLS; Rajah & McIntosh, 2008). Behavioural results showed that older adults performed as well as young on item recognition, but performed worse than young on both context tasks. The PLS analysis identified a significant latent variable (LV) representing a task main effect of context versus recognition tasks at retrieval, but not at encoding, in both age groups. Specifically, both groups activated left dorsolateral PFC (Brodmann area [BA] 9), left BA 47 and left parahippocampal gyrus to a greater degree during context versus item retrieval. We found a significant LV representing group by task interactions within retrieval, but not encoding tasks. Specifically, left hippocampus was preferentially used by the young during context retrieval, while the elderly used it preferentially during recognition. Finally, we observed group differences in left BA 44 during encoding (all tasks) versus retrieval (all tasks). Young activated this region preferentially at encoding, and the elderly at retrieval.

Slide Session 5
Monday, April 19, 3:00 - 5:00 pm, Outremont et al Ballroom

Working Memory and Executive Functions

Chair: Silvia Bunge, University of California, Berkeley

Presentation 1 (GSP Winner): The capacity limit of visual working memory in the macaque monkey

Evelien Heyselaar¹, Kevin Johnston¹, Martin Paré¹; ¹Queen's University Canada

Working memory is a limited capacity system that allows the temporary retention of information to guide future behavior. The human visual working memory capacity has been estimated using the sequential comparison task. This limit is approximately 3 items, with values as low as 1.5 in some individuals. However, similar limits have not yet been systemically investigated in other animals, which is necessary toward development of an animal model for investigation of the neural basis of this facet of cognitive function. We estimated the visual working memory capacity of the macaque monkey using the same procedure as human studies. In this task, the animal was first presented with a memory array consisting of a varying number (set size) of highly discriminable color stimuli. The offset of the memory array was followed by a retention interval, during which the animal was required to maintain fixation. Following this, a test array was presented in which the color of one of the stimuli had changed. The animal was required to identify which stimulus had changed by making a single saccadic eye movement to its location. Animals’ ability to detect a color change declined with increases in the set size and duration over which they were retained in working memory. From the relationship between performance and set size, we estimated that visual working memory in the macaque monkey has a capacity limit that exceeds two memoranda. This limit is comparable to that of humans, suggesting a shared neural substrate for this aspect of executive function.

Presentation 2: Dissociation within the fronto-parietal network in verbal working memory: a parametric fMRI study

Anne Sophie Champod¹, Michael Petrides¹; ¹Montreal Neurological Institute, McGill University

Concomitant increase in activity within the mid-dorsolateral prefrontal cortex (MDLFC) and the posterior parietal cortex (PPC) is observed in most functional neuroimaging studies of working memory. Despite broad consensus on the importance of these two brain regions in working memory, the unique contribution of the PPC remains a matter of heated debate. The main objective of the present parametric event-related functional magnetic resonance imaging (fMRI) study was to examine the hypothesis that the cortex in the intraparietal sulcal (IPS) region in the PPC is involved in the manipulation (i.e. rearrangement) of verbal information in working memory and to dissociate the involvement of this brain region from the known involvement of the MDLFC in the monitoring of information in working memory. The results demonstrated a linear increase in activity within the MDLFC during the manipulation and monitoring of a linearly increasing number of words in working memory. In sharp contrast, there was a linear increase in activity within the PPC during the manipulation, but not the monitoring of a linearly increasing number of words. This study provides the first parametric dissociation of activation in these two cortical regions, indicating a crucial role of the PPC in the manipulation of information in working memory, with the MDLFC playing a major role in monitoring this information.

Presentation 3: Functional connectivity between networks underlying working memory is modulated by dopamine-regulating genes COMT and DAT

Evan Gordon¹, Melanie Stollstorff², Priyanka Salona², Chandan Vaidya^2,3; ¹Georgetown University Medical Center, ²Georgetown University, ³Children's National Medical Center

Functional magnetic resonance imaging (fMRI) reveals that some regions are activated (task-positive network - TPN) whereas others are deactivated (default mode network - DMN) during working memory (WM) performance. Studies have reported that these networks are negatively correlated; however, the nature of DMN-TPN connectivity varies across individuals. We examined whether DMN-TPN connectivity differed by genotype for the catechol-O-methyltransferase (COMT) and dopamine transporter (DAT) genes, which induce individual variation in synaptic levels of dopamine. FMRI was performed on 40 subjects during rest and during performance of a WM task. For each subject, we identified peak voxels of deactivation during WM within the posterior cingulate cortex (PCC), a DMN region, and used those as seeds for functional connectivity analysis after regressing out signal from ventricles. We observed a COMTxDAT interaction in PCC connectivity with right dorsolateral prefrontal cortex (a TPN region) during WM. Post-hoc analyses indicated that only 9/10 Met/Met subjects demonstrated negative connectivity, suggesting that differences in dopamine levels modulated the direction of the DMN-TPN relationship. We subsequently examined whether the individual differences in the DMN-TPN relationship observed during WM were also present at rest. The COMTxDAT interaction was not observed during rest, and furthermore, DMN-TPN connectivity during WM only correlated with connectivity during rest after accounting for effects of genotype on WM connectivity (residual r=.459, p<.005). This result suggests that “intrinsic” DMN-TPN connectivity observed during rest is altered due to effects of dopamine release during WM, which is affected in turn by COMT and DAT genotype.

Presentation 4: Changes in distributed patterns of neural activation associated with proactive interference resolution in working memory

Ilke Öztekin¹, David Badre¹; ¹Brown University

An obstacle to successful retrieval is the presence of interference during the retrieval attempt. Proactive interference (PI), which arises from a prior learning event, is a major cause of forgetting. In this study, we sought to identify distributed patterns of neural activity that mark the retrieval of irrelevant information. This irrelevant information causes PI and leads to forgetting in working memory. Participants were scanned using functional magnetic resonance imaging (fMRI) while performing a short-term item recognition task. Trials consisted of a 5-item study list and then a recognition probe following a distractor period. To induce PI, three consecutive study lists were constructed from the same semantic category (e.g., animals), and then the category was switched (e.g., fruits) on the subsequent trial. Recognition performance decreased across successive lists of the same category, consistent with the buildup of PI. We employed a multi-voxel pattern classifier approach to track distributed patterns of brain activity that mark the retrieval of each category. Classifier success indexed the strength of the representation of each category in the brain, independent of behavioral success. Classifier success for the common category increased across the lists, consistent with the category information becoming increasingly elicited during retrieval. Critically, this increase was correlated with behavioral measures of forgetting, suggesting that the representation of non-diagnostic category information leads to a decline in memory performance. Conversely, classifier success was negatively correlated with neural activation in the left ventrolateral prefrontal cortex, consistent with this region’s hypothesized role in successfully resolving competition during memory retrieval.

Presentation 5: More GABA, less distraction: A neurochemical correlate of variability in human eye movement control

Petroc Sumner¹, Richard Edden², Aline Bompas¹, Krish Singh¹; ¹Cardiff University, ²Johns Hopkins University

Every researcher of human behaviour knows that people perform differently, and this remains true even in the most basic situations involving the fundamentals of simple actions. Yet, although stimulating growing interest, the reasons for such differences are generally unknown. Here we show that differences in basic performance can correlate tightly with subtle neurochemical differences in relevant brain regions. We investigate the ability to inhibit an irrelevant distractor when making eye movements, and find that it is well predicted by the concentration of GABA – the main inhibitory neurotransmitter – measured by magnetic resonance spectroscopy in a region including human frontal eye field (FEF), but not in a control region. Moreover, by employing a model that distinguishes three types of motor inhibition, we can specifically associate GABA variation in the FEF region with just one of them – top-down reactive inhibition.

Presentation 6: Dissociable contributions of human prefrontal cortex in dynamically updating behavior

Frederick Verbruggen¹, Adam Aron², Michael Stevens¹, Christopher Chambers³; ¹Ghent University, ²UCSD, ³Cardiff University

The human cognitive control system supports flexible and goal-directed behavior by updating actions in response to changes in the environment. Cognitive psychologists have shown that dynamic updating of behavior depends on fast-acting control mechanisms that stop or change responses when they become irrelevant or inappropriate. It is currently intensely debated which regions in human prefrontal cortex support stopping and updating of behavior, and competing accounts need to be reconciled. We investigated the functional specificity of lateral prefrontal cortex in updating behavior by combining transcranial magnetic stimulation (TMS) with a novel behavioral paradigm, in which subjects had to stop or update responses on a minority of the trials. The results show that one region of lateral prefrontal cortex, the right inferior frontal junction is crucial for updating the attentional focus when an infrequent signal occurs, whereas a different region, the right inferior frontal gyrus is crucial for updating action plans in response to this signal. These findings shed new light on the role of lateral prefrontal cortex in stopping and updating behavior. In particular, the idea that right inferior frontal gyrus implements control by updating action plans reconciles recent controversial and conflicting findings about its functional role.

Presentation 7: Theta burst stimulation over the pre-SMA improved response inhibition on a conditional stop signal reaction time task

Ignacio Obeso¹, Leonora Wilkinson¹, Teo James¹, Talleli Penelope¹, John Rothwell¹, Marjan Jahanshahi¹; ¹UCL Institute of Neurology, London,UK

Introduction: Volitional control over our actions often necessitates inhibition of ongoing responses. This has been investigated with the stop signal reaction time task (SSRTT). Imaging showed activation of right inferior cortex (IFC), pre-supplementary motor area (pre-SMA) and subthalamic nucleus (STN) during the SSRTT (Aron et al, 2007; Li et al, 2008). There is no consensus about their specific roles, some suggesting that the IFC and STN are involved in inhibition, while pre-SMA mediates conflict monitoring/resolution (Aron et al, 2007). Others attribute an inhibitory role to both the IFC and pre-SMA (Li et, 2008). We used the conditional SSRTT (cSSRTT) and measured inhibition and conflict induced slowing to determine if continuous theta burst stimulation (cTBS) over the IFC and pre-SMA would produce similar or dissociable effects on these measures. Methods: In a within-subjects design, fourteen volunteers performed the cSSRTT on three days (weekly intervals) after ‘offline’ continuous cTBS over the right IFC, pre-SMA or sham M1, with order randomized. We measured the stop signal reaction time (SSRT) and conflict induced slowing (CIS). Results: Relative to sham TBS, cTBS over pre-SMA resulted in significantly shorter SSRTs. This effect was specific to the SSRT and not observed for CIS or any other RT measures or errors. cTBS over the right IFC had no effect on any RT or error measures. Discussion: The unexpected significant shortening of SSRT with cTBS over the pre-SMA suggests that the pre-SMA made a critical contribution to volitional inhibition on the cSSRTT but did not influence conflict monitoring/resolution.

Presentation 8: Adolescent executive function and brain structure following childhood traumatic brain injury

Miriam Beauchamp^1,2, Cathy Catroppa^2,3,4, Celia Godfrey², Sue Morse³, Jeffery Rosenfeld^5,6, Vicki Anderson^2,3,4; ¹Department of Psychology, University of Montreal, Canada, ²Murdoch Children's Research Institute, Melbourne, ³Royal Children's Hospital, Melbourne, ⁴School of Behavioural Sciences, University of Melbourne, ⁵Department of Surgery, Monash University, ⁶Department of Neurosurgery, The Alfred Hospital, Melbourne

The impact of brain injury on skills that allow individuals to engage in purposeful, goal-directed and appropriate behaviour, commonly conceptualized as executive functions, has been a focal point of research into the cognitive outcomes of paediatric traumatic brain injury (TBI). However, few studies have addressed the long-term evolution of executive deficits in patients injured during childhood. Forty children with TBI were followed longitudinally and assessed ten years post-injury (25 male, mean age at injury = 2.0, SD = 2.7, mean age at follow-up = 14.7, SD = 0.7) and compared to nineteen healthy developing children (12 male, mean age at follow-up = 14.2, SD = 2.3) on a battery of tests measuring a range of executive functions including attentional control, inhibition, cognitive flexibility, planning, goal setting, problems solving and abstract reasoning. Results revealed that patients with mild and moderate injuries performed at age-expected levels on all tests; however severely injured patients demonstrated significant deficits particularly on complex executive functions such as planning [(F(2, 54) = 3.3, p =.02)], problem solving [F(3, 55) = 3.4, p = .02], and abstract reasoning [F(48, 3) = 5.1, p = .004]. These findings highlight the lasting effects of childhood brain injury on complex executive skills and suggest that severely injured children may require ongoing intervention and support for cognitive deficits into adolescence and early adulthood.

Slide Session 6
Tuesday, April 20, 10:00 am - 12:00 pm, Outremont et al Ballroom

Attention

Chair: Marty Woldorff, Duke University

Presentation 1: Competition for cognitive resources during rapid serial processing: changes across childhood

Sabine Heim¹, Nadine Wirth², Andreas Keil³; ¹German Institute for International Educational Research (DIPF), ²Central Institute of Mental Health, Mannheim, ³University of Florida

Changes in the strategic control of attentional resource allocation play an important role for the development of complex cognitive skills and academic achievement. We examined such changes in a cross-sectional design, using the attentional-blink (AB) paradigm. The AB is an impairment of T2 report, which occurs when a first (T1) and second target (T2) embedded in a rapid stimulus sequence appear within 500 ms of each other. Two groups of children (6 to 7 year-olds and 10 to 11 year-olds; ns = 21 and 24, respectively) worked on two AB tasks (symbols in experiment 1; letters or words in experiment 2), identifying green targets in an 8.7 Hz stream. The temporal distance between T1 and T2 varied between no intervening distractor (lag 1) up to 7 intervening distractors (lag 8). Both experiments showed that older children performed more accurately overall. In the symbol task, younger children linearly increased T2 performance with increasing lag, whereas older children displayed a quadratic pattern as typically seen in adults, with a slight performance gain for the lag-1 condition. In the verbal task, younger individuals again exhibited cost effects for lag 1, whereas the older showed sparing. Impairment of T2 report at lag 1 in the younger children was not related to complementing T1 effects: they displayed greater T1 performance problems at lag 1 than older participants. Taken together, this pattern of results suggests that strategic control of temporal attention resources is a skill that emerges during primary school age.

Presentation 2 (GSP Winner): Visual Working Memory Mediates Visual Search Efficiency

Stephen M. Emrich¹, Naseem Al-Aidroos¹, Jay Pratt¹, Susanne Ferber^1,2; ¹University of Toronto, ²Rotman Research Institute, Baycrest

Visual search requires the maintenance of previously searched distractors to bias selection towards novel items. It is unclear, however, which memory system supports the indexing of these old items. Interestingly, tasks that depend on visual working memory (VWM), a capacity-limited short-term visual store, elicit an event-related potential (ERP) contralateral to the attended memory items known as the contralateral delay activity (CDA). Accordingly, in the present study we tested whether VWM supports the inhibition of old items by testing for the presence of the CDA during visual search. Our results demonstrate that a lateralized search task elicited activity contralateral to the attended search array that was indistinguishable in mean and peak amplitude from the CDA observed during a four-item change-detection task. The change in amplitude of this contralateral search activity (CSA) over time was strongly correlated with VWM capacity, suggesting that this activity reflects VWM processing. Furthermore, behavioral measures of search efficiency were strongly correlated with changes in CSA amplitude, as well as with an independent measure of VWM capacity. That is, the more items that could be stored in memory, the faster the search target was found. These findings indicate that VWM likely supports the inhibition of previously searched distractors, and consequently, that visual search efficiency is closely linked to individual VWM capacity.

Presentation 3: Neuronal oscillations as a tool for attentional selection in the cocktail party effect: insights from intracranial EEG in humans

Elana Zion Golumbic^1,2, Julien Besle¹, Ashesh Mehta³, Catherine Schevon¹, Ronald Emerson¹, David Poeppel⁴, Charles Schroeder^1,2; ¹Columbia University, ²Nathan Klein Institute for Psychiatric Research, New York, ³Long Island Jewish Hostpital, New York, ⁴New York University

Our ability to attend to a particular conversation amidst competing input stream, epitomized by the “cocktail party” effect, is remarkable. Yet, the neural mechanisms involved in segregating, selecting and attending to one stream of conversation are not well understood. Here we tested the idea that such selection is brought about by adjusting the phase of ongoing neuronal oscillations in auditory cortex to entrain to the temporal envelope of the attended speech stream. We recorded intracranial EEG in patients undergoing electrode implantation for clinical purposes, while participants listened to segments of natural speech. We simulated a “cocktail party” by playing two conversations concurrently (male and female speakers) and instructed participants on each trial to attend to one speaker and ignore the other. In a control block the same speech segments were presented alone. In each condition, we analyzed the consistency of phases across trials (phase-locking, PL) between 1-50Hz in electrodes over auditory cortex. In the cocktail-party conditions, despite the identical physical stimulus, the PL pattern was modulated as a function of attention. Specifically, the PL pattern elicited when attending to a particular segment was highly correlated with the PL pattern elicited when listening to that segment alone (in the control condition). In addition, similarities were observed between the PL pattern and the spectro-temporal structure of the speech envelope. The results strongly support the idea that the brain entrains its internal neuronal oscillations to the rhythmic structure of an input, and that this mechanism is utilized as a tool for selective attention.

Presentation 4: Adaptive representation of task-relevant information in frontoparietal cortex

Alexandra Woolgar¹, John Duncan¹; ¹Medical Research Council - Cognition and Brain Sciences Unit

In human functional magnetic resonance imaging (fMRI), a characteristic pattern of frontal and parietal activity is produced by many different cognitive demands. These ‘multiple-demand’ regions are thought to be critical for flexible goal-directed behavior. The adaptive coding hypothesis posits that they achieve cognitive control by the adaptive representation of task-relevant information. We used multi-voxel pattern analysis (MVPA) of human fMRI data to demonstrate adaptive coding of stimulus, rule and response information in a visual stimulus-response task. When the perceptual difficulty of the task was increased, frontoparietal regions showed significantly increased coding of perceptual information. At the same time, low-level visual areas showed significantly weakened coding of perceptual information, in line with the physical change in the visual stimulus. These changes occurred dynamically, across short alternating blocks of high and low perceptual difficulty. Coding of other task-features was concurrently redistributed across the multiple-demand system, such that together the frontoparietal regions continued to represent all the information needed for the task. The results are consistent with the conception of the frontoparietal multiple-demand regions as a flexible neural system that exerts cognitive control in a wide range of tasks by adaptively representing the task-features that are relevant for successful goal-directed behavior.

Presentation 5: Areas underlying the deployment of spatial attention in the human brain: A ¨blind¨ TMS mapping of parietal and frontal regions

Antoni Valero-Cabre^1,2, Claudia Peschke³, Kurtev Stoyan³, Roxana Voitcu³, Yu Jin³, Richard Rushmore², Claus Hilgetag^3,4; ¹CNRS UMR 5105-INSERM S975, Paris, France, ²Boston University School of Medicine, ³Jacobs University Bremen, Germany, ⁴Boston University

Prior TMS studies have suggested that a myriad of distinct regions in the human brain contribute to the processing of spatial attention. We used a basic visual detection/localization task and applied TMS to systematically map, in an anatomically “blind” manner, the right and left parietal and frontal subregions likely to generate significant visuo-spatial biases. Two groups of 8 adult subjects executed a task based on the detection of black dots on light-gray background. Targets were displayed unilaterally or bilaterally for 40 ms, at 25o with a back-projection system. In two independent experiments, a 9 (3x3) stimulation sites grid was anchored lateral to P3 and P4 EEG coordinates for the parietal mapping and 2 cm rostral to M1 for the frontal mapping. Three pulses of real or sham 10 Hz TMS were delivered on each grid location 50 ms post target onset to interfere with the ongoing stimulus processing. In the parietal mapping, real TMS on the right and left IPS region, 2 cm ventral to IPS, and slightly above TPJ induced mirror-symmetric ipsilateral spatial biases. In the frontal mapping, significant reductions in detection rates for contralateral stimuli and increased detection performance for ipsilateral targets were observed for the region anatomically associated with right FEF, while TMS of left frontal sites yielded spatially non-selective results. We conclude that the effects of TMS on a simple spatial localization task for which a well-balanced deployment of attention is required, are exquisitely spatially selective, and mostly noted in specific right parietal and frontal regions.

Presentation 6: The causal roles of the dorsal and ventral fronto-parietal networks in orienting spatial attention

Ana Chica¹, Antoni Valero^1,2, Paolo Bartolomeo^1,3,4; ¹INSERM-UPMC UMR-S 975, Paris, France, ²CNRS-UMR 5105, Paris, France, ³AP-HP, Groupe Hospitalier Pitié-Salpêtrière, Fédération de Neurologie, Paris, France, ⁴Catholic University, Milan, Italy

Attention can be spatially oriented in a top-down or bottom-up manner. Corbetta et al. (2008) have proposed that a dorsal network (including the Intra-Parietal Sulcus –IPS, and the Frontal Eye Field –FEF) is mostly involved in attentional orienting (both bottom-up and top-down), while a ventral network (including the Temporo-Parietal Junction –TPJ, and the Inferior Frontal Gyrus –IFG) is not related to the orienting of attention but is key in re-orienting attention to unexpected events. We tested the causal roles of the right IPS and TPJ during bottom-up orienting of spatial attention using Transcranial Magnetic Stimulation (TMS). Participants were presented with a spatially non-informative peripheral cue that preceded the to-be-discriminated target by either 200 or 800 ms. Two TMS pulses were delivered 50 ms post-cue presentation to interfere within a 100 ms time window. Attentional effects for cues presented ipsilaterally to the stimulation side were as expected, showing facilitation at the short interval, and inhibition at the longer one. Importantly, for contralateral cues, a dissociation was observed when TMS was applied to IPS and TPJ. TMS pulses on IPS abolished the orienting of attention while TMS on TPJ produced a facilitatory effect even at the long SOA, where inhibition was expected. Our results indicate that both the dorsal and the ventral network play a causal role in orienting attention, although their effects are clearly dissociable. The abnormal facilitation observed after TPJ stimulation mimics the characteristic attentional deficits observed in patients with right hemisphere damage suffering from neglect (Bartolomeo & Chokron, 2002).

Presentation 7: Neural correlates of attentional bias toward cigarette cues in active smokers

Vicki Chanon¹, Charlotte Boettiger¹; ¹University of North Carolina at Chapel Hill

Data suggest that craving triggered by drug cue exposure results in part from abnormal allocation of attention to drug cues in substance-dependent individuals (SDIs). The importance of investigating the neural bases of such bias is highlighted by evidence that its strength predicts risk of relapse. Little is known about the neural correlates of addiction-related attentional bias. To address this issue, we tested active smokers (AS, n=12) and non-smokers (NS, n=11) using fMRI with a spatial cuing paradigm designed to measure attentional bias toward smoking-related images. Behavioral measures revealed a bias toward smoking cues in AS. We also observed selective enhancement of activity in the Inferior Frontal Gyrus in response to smoking cues, specific to the AS group. We speculate that this reflects a need for increased inhibitory control in AS when the salient, yet irrelevant smoking cues were present. In addition, we observed enhanced activity, specific to AS, in response to targets congruent with a smoking cue (relative to incongruent) in a constellation of areas implicated in attention and cognitive control, including the inferior frontal junction, orbitofrontal cortex (OFC), anterior cingulate cortex (ACC), temporoparietal junction, and precuneus, as well as the right caudal hippocampus. Abnormalities in the OFC and ACC are implicated in executive dysfunction among SDIs. Moreover, the ACC projects throughout the frontoparietal attention system. Our data suggest that the ACC biases this system to attend to smoking stimuli. Identifying these biomarkers of addiction-related attentional bias provides new putative targets for development of therapeutic interventions for addictive disorders.

Presentation 8: Nicotinic receptor gene CHRNA4 interacts with processing load in attention

Thomas Espeseth¹, Markus Sneve¹, Helge Rootwelt², Bruno Laeng¹; ¹University of Oslo, Norway, ²University Hospital of Oslo, Norway

Pharmacological studies have suggested that cholinergic neurotransmission mediates increases in attentional effort in response to high processing load in attention demanding tasks (Sarter et al. 2006). In the present study we tested whether individual variation in CHRNA4, a gene coding for a subcomponents in the alpha4beta2 nicotinic receptors in the human brain, interacted with processing load in a multiple object tracking task (MOT) and a visual search task (VS). We hypothesized that genotype would predict performance in both tasks under high load, but not low load. One hundred forty six healthy persons (age range=39 – 77, mean=55.6, sd=9.0) performed the MOT task in which twelve identical circular objects moved about the display in a random manner. Two to six objects were targets and the remaining objects were distractors. The same observers also performed a visual search for a target letter (i.e. X or Z) presented together with five non-targets while ignoring centrally presented distractors (i.e. X, Z, or L). Targets differed from non-targets by a unique feature in the low load condition, whereas they shared features in the high load condition. CHRNA4 genotype interacted with processing load in both tasks. Homozygotes for the T allele (N=48) had better tracking capacity in the MOT task and identified targets faster in the VS task, and these differences occurred only in high load conditions. The results support the hypothesis that the cholinergic system modulates attentional effort, and that common genetic variation can be used to study the molecular biology of cognition.

Slide Session 7
Tuesday, April 20, 1:00 - 3:00 pm, Outremont et al Ballroom

Perception & Action

Chair: Kevin Wilson, Gettysburg College

Presentation 1: The relationship between speech perception and the human "mirror system": evidence from lesion case-studies

Corianne Rogalsky¹, Tracy Love², David Driscoll³, Steven Anderson³, Gregory Hickok⁴; ¹University of Southern California, ²San Diego State University / University of California, San Diego, ³University of Iowa, ⁴University of California, Irvine

The discovery of mirror neurons in macaque has led to a resurrection of motor theories of speech perception. In particular, it has been proposed that the “human mirror system”, which prominently includes Broca’s area, is the neurophysiological substrate of speech perception. Although numerous studies have demonstrated a tight link between sensory and motor speech processes, few have directly assessed the critical prediction of mirror neuron theories of speech perception, namely that damage to the human mirror system should cause severe deficits in speech perception. The present study tested this prediction by measuring speech perception abilities of subjects with lesions involving motor regions in the posterior frontal lobe and/or the left inferior parietal lobule (i.e. the proposed human "mirror system"). A subset of subjects had lesions that also included temporal regions. Subjects completed a psycholinguistic battery to assess their phonological, lexical, and sentence-level speech comprehension and production abilities. Signal detection analyses indicate that performance was near ceiling in the word comprehension and discrimination tasks for patients with damage to the “human mirror system". Perceptual deficits were evident only when the lesion encroached on auditory regions in the temporal lobe. Perceptual deficits in patients with temporal lobe involvement were particularly evident on non-word stimuli compared to word stimuli even when stimuli were equated for phonological neighborhood density and phonotactic frequency. These results suggest that damage to the human "mirror system" does not disrupt speech perception and argue instead that auditory systems are the primary substrate for speech perception. Supported by NIH-DC03681.

Presentation 2 (GSP Winner): Abstract coding of movement direction in human superior parietal cortex

Sara Fabbri¹, Alfonso Caramazza^1,2, Angelika Lingnau¹; ¹University of Trento, ²Harvard University

Neurons in monkey motor cortex are broadly tuned for movement direction: activity of a neuron is highest for the preferred movement direction and decreases gradually with increasing angular difference between the preferred and non-preferred direction. Here we aimed to identify areas with similar properties in the human brain. In particular, we asked if there exists an abstract representation of movement direction that remains invariant despite changes in low-level parameters, like the type of motor act (to grasp vs to press). Using functional magnetic resonance imaging (fMRI) adaptation, we identified several areas within the human visuomotor system whose blood-oxygen level dependent (BOLD) activity was modulated by the angular difference between adapted and test hand movement directions. In particular, right superior parietal lobe (SPL) showed pronounced directional tuning that was not affected by the type of motor act (to press vs to grasp). Most other regions in the dorsal pathway showed directional tuning that was significantly modulated by the type of motor act. In contrast, the BOLD response in left and right cerebellum was modulated by movement direction and type of motor act, but not by their interaction. These results suggest that the control of directional movements in the human brain relies on a network of areas selective for movement direction at different levels of abstractness. Such an abstract representation of movement direction might play an important role for the dynamic control of movement direction irrespective of changes in low-level muscle parameters as well as for transformations from visual to motor coordinates.

Presentation 3: MEG signatures of saccadic choice and speed differ: a critical test for decision models

Aline Bompas¹, Iain Gilchrist², Suresh Muthukumaraswamy¹, Krish Singh¹, Petroc Sumner¹; ¹Cardiff University, CUBRIC - School of Psychology, ²Bristol University, Psychology Department

A fundamental prediction of all accumulator models of decision is that choice and latency are related, because latency to one visual target and choice between two possible targets are determined by the same factor – how soon activity for a particular response reaches threshold in an action planning area. We used magnetoencephalography (MEG) to design a critical test for this attractive though unchallenged assumption. Participants made saccades to onset targets (left or right), and on one third of the trials had to choose freely between two simultaneous targets (left and right). We split these bilateral trials according to the choice made (left versus right) and compared the MEG activity during the 200 ms before the stimulus onset (baseline). Similarly in single target trials, we compared the baseline MEG activity preceding fast and slow saccades. While choice was predicted by frontal differences in the alpha frequency band (5-15Hz), speed was predicted by non-lateralised occipital differences in the alpha and beta bands (5-30Hz), the two signatures showing no similarities, even at the sub-significant level. Our results suggest that different baseline settings specifically modulate choice and speed, in contrast with the central assumption of accumulator models.

Presentation 4: Modality-specific and modality-independent components of the human imagery system

Sander Daselaar¹, Yuval Porat¹, Willem Huijbers¹, Cyriel Pennartz¹; ¹University of Amsterdam

Imagery research typically deals with the commonalities and differences between imagery and perception. As such, it is usually confined to one specific modality. Yet, it is likely that some of the underlying processes are shared between different sensory modalities while others are modality-specific. In this fMRI study, we used a balanced design that allowed for a direct comparison between imagery and perception in visual and auditory modalities, and also for a link between subjective imagery experience and brain activation. Results indicated a selective role for the “default mode network” as a modality-independent “core” imagery network. This finding supports the idea of a general role of the DMN in rich mental imagery involving multiple sensory components. In addition, results identified areas in the visual and auditory association cortices that contributed to mental imagery in a modality-specific fashion. The finding of shared imagery-related and perception-related activity in VA and AA within one single task suggests that, even though these regions are processing different features of imagined objects, they share a common processing level situated at an advanced stage in the hierarchy of the human sensory system. Interestingly during mental imagery, primary visual and auditory cortices showed modality-specific suppression of activity. Suppression of primary sensory regions may help the processing of internally-generated images or sounds by shielding the associative sensory regions from external perceptual input processed by primary regions. This is the first fMRI study to characterize both modality-specific and modality-independent components of the human imagery system.

Presentation 5: Manipulating Visual Perception with Real-Time fMRI-based Neurofeedback Training

Frank Scharnowski¹, Chloe Hutton¹, Oliver Josephs¹, Nikolaus Weiskopf¹, Geraint Rees¹; ¹University College London (UCL)

Spontaneous fluctuations of ongoing brain activity have a profound impact on perception. For example, human observers are more likely to perceive a visual stimulus if baseline activation in visual cortex at the time of presentation is higher. Here, we used real-time fMRI-based neurofeedback to train participants to voluntarily regulate the ongoing brain activation of circumscribed areas in their early visual cortex. We then tested how such self-regulated activation influenced subsequent processing of a visual stimulus presented near threshold. We found that the level of activation within early visual cortex directly influenced objective detection thresholds, i.e. when the participants increased activation they became better at detecting a visual stimulus. A control group who received feedback from a non-visual brain region did not learn to control visual cortex activation and did not show changes in visual sensitivity. Hence, with real-time fMRI-based neurofeedback it is possible to learn voluntary control over visual cortex activation and thereby to improve visual sensitivity. This new approach allows us to now study perception as a variable which is dependent on self-regulated brain activation, and therefore to investigate causal links between brain activation and perception.

Presentation 6: Neural Correlates of Short-Term Visual Plasticity and Cortical Disinhibition in Humans

Nathan Parks¹, Paul Corballis²; ¹Beckman Institute, University of Illinois at Urbana-Champaign, ²Georgia Institute of Technology

Restricting visual input from a circumscribed region of space induces representational plasticity within retinotopically organized visual areas such that deafferented cortex begins to represent new spatial locations. The extent of such plasticity is most compelling over the long-term but functional changes begin within seconds of retinal deafferentation. The neural mechanism proposed to underlie short-term visual plasticity is one of disinhibition whereby a reduction of local inhibition within deafferented areas renders them more sensitive to inputs from surrounding cortex. We investigated the neural mechanisms of short-term visual plasticity and disinhibition in humans using a stimulus-induced analog of retinal deafferentation known as an artificial scotoma – a uniform area stabilized upon a dynamically changing background. Artificial scotomas provide a useful model of short-term visual plasticity as they temporarily mimic the perceptual and neural changes that accompany a true retinal scotoma (pseudo-deafferentation). In a series of studies we measured contrast response functions and visual evoked potentials (VEPs) from within the boundaries of an artificial scotoma (scotoma condition). In comparison to a closely matched control (sham condition), psychophysical functions exhibited contrast-dependent effects indicative of increased spatial integration but decreased stimulus selectivity. VEPs elicited from within the boundaries of an artificial scotoma exhibited enhanced amplitude of early sensory components, indicating that reduced stimulus selectivity is associated with increased neural gain in pseudo-deafferented visual cortical areas. Results follow the predictions of disinhibition and suggest that short-term visual cortical deafferentation is associated with increased spatial integration as well as a nonselective increase in neural response gain.

Presentation 7: The Cognitive, Functional, and Structural Architecture of Cross-Hemispheric Communication in Younger and Older Adults

Simon Davis¹, James Kragel¹, David Madden¹, Roberto Cabeza¹; ¹Duke University

In order to investigate the neural mechanisms of cross-hemispheric communication and how they change as a function of aging, we (1) assessed cross-hemispheric transfer using bilateral-unilateral matching task, (2) linked behavioral results in this task to fMRI connectivity (FC) between homologous regions across hemispheres, and (3) linked FC to DTI measures of white matter integrity in the corresponding cross-hemispheric tracts. (1) Replicating previous research, we observed that (a) matching was faster for bilateral than unilateral trials (bilateral processing advantage), (b) that this effect increased with matching difficulty, and (c) occurred a lower difficulty levels for older than younger adults. (2) FC was greater at higher levels of task difficulty between homologous regions in frontal, temporal, and parietal lobes; older adults showed a similar pattern but in a greater number of homologous regions. (3) Finally, FC in several pairs of homologous regions was significantly correlated with white matter integrity (fractional anisotropy) in the corresponding tracts; white matter integrity was also correlated with the bilateral processing advantage. These effects were stronger in older than younger adults. Taken together, the findings support a hypothetical causal chain whereby changes in white matter integrity (#3) modulate changes in FC (#2) which in turn modulate cross-hemispheric transfer and behavioral performance (#1). The results are also consistent with the notion that greater communication between the hemispheres engenders more efficient processing in response to increased task demands, and that older adults show greater reliance on a bilaterally distributed network, possibly reflecting functional compensation.

Presentation 8: Multi-voxel similarity structure analysis reveals taxonomy of animal species in human ventral temporal cortex

Andrew C. Connolly¹, James V. Haxby¹; ¹Dartmouth College

Our goal is to develop a detailed account of the structure of the representation of living things in the brain and how this structure emerges in the object vision pathway, investigating three regions: medial occipital (MO), inferior occipital (IO), and ventral temporal cortex (VT). Our approach is based on the analysis of similarity structure for multi-voxel patterns defined by responses to a variety of animate categories using functional magnetic resonance imaging (fMRI). During fMRI scanning, subjects (N=12) viewed photographs of six animal species—two species each of insects, birds, and primates. Pair-wise distances between condition patterns were used to construct 15-dimensional neural similarity spaces, where each dimension corresponds to the distance between a species pair. The similarity structures revealed how categorical representations take shape along the visual processing pathway. Patterns in early visual cortex (MO) are less differentiated in general than in IO and VT, and without clear category structure. In IO, there is strong differentiation between the vertebrates and the insects, while in VT each intermediate-level category becomes clearly defined. Similarity structures are highly stable and replicable both within and between subjects—especially in VT with an average between-subject correlation of r=.91, which is higher than that for regions earlier in the visual processing stream: IO, r=.75; MO, r=.65. Similarity-based multi-voxel pattern analysis reveals a categorical structure in ventral temporal cortex that mirrors our knowledge about animal species, providing a window into the structure of neural representations that form the basis of our categorical knowledge of the living world.