Noon - 1:30 pm
Noon - 1:30 pm
Data Blitz Session 1
Saturday, March 25, Noon - 1:30 pm, Bayview
Chair: Marian Berryhill, University of Nevada
Speakers: Yuri Dabaghian, Ryan Giuliano, Anna McCarrey, Alessandro Tavano, Anna Magdalena Barth, Elizabeth L. Johnson, Kevin Jones, Zhang Jingting, Heather D. Lucas, Milena Rabovsky, Anna Khazenzon, Matthew Sazma, Layla Unger, Joe Bathelt, Pedro Pinheiro-Chagas
Talk 1: Internal consistency of spatial information in a cognitive map
Yuri Dabaghian1; 1Baylor College of Medicine, Houston, TX 77019 USA
Learning and memory emerge from the activity of groups of neurons, yet there are few models that can meaningfully connect data acquired at the level of individual cells with whole-animal behavior beyond mere correlation. We have begun to tackle this problem by computationally modeling spatial learning. The foundation of our approach is the hypothesis that the hippocampus provides a rough-and-ready topological framework of an environment rather than a precise metrical map. This model, which is supported by animal studies, allows us to employ algebraic topology to ascertain the effects of specific parameters (e.g., firing rate) on the ability of an ensemble of virtual neurons to correctly “learn” an experimental environment. We have recently studied the effects of two brain rhythms, θ- and γ-waves, on spatial learning. Both have been correlated with learning but it has been difficult to explain precisely why. We found that θ-phase precession parcellates place cell coactivity at the network scale (~150-200 msec), as recorded in animal experiments, and show how this enhances spatial learning. We also found that γ-rhythm synchronizes spiking in dynamical place cell assemblies to enable encoding and retrieval of spatial memories at the synaptic timescale (~50 msec). Topological theory thus provides a conceptually elegant description of the spatial learning process and enables us to explain a wide range of phenomena.
Talk 2: Cardiac Measures of Autonomic Arousal are Associated with ERP Measures of Selective Attention in Children and Adults
Ryan Giuliano1, Christina Karns1, Theodore Bell1, Leslie Roos1, Seth Petersen1, Elizabeth Skowron1, Helen Neville1, Eric Pakulak1; 1University of Oregon
Neurovisceral integration theory stipulates that higher-order functions of the brain, in particular those indexing networks involving the prefrontal cortex, are intricately linked to the regulation of autonomic physiology. However, few studies include simultaneous neural and autonomic measures. To this end, we recruited young children and adults for a laboratory visit where we recorded cardiovascular measures of parasympathetic and sympathetic nervous system activity, respiratory sinus arrhythmia (RSA) and pre-ejection period (PEP) respectively, during an ERP dichotic listening measure of auditory selective attention. During the dichotic listening task, participants were simultaneously presented with two narrators reading different stories from speakers to their left and right, and ERPs were examined as mean amplitudes evoked by auditory stimuli embedded in stories they were asked to attend versus stories they were asked to ignore. Results demonstrated that for both children and adults, cardiac arousal was associated with better indices of selective attention. Among children, ERP amplitudes elicited by distracting stimuli were inversely associated with arousal, such that children with the smallest ERP response to distractor sounds also had the shortest resting PEP (i.e., higher sympathetic activity) and showed the greatest RSA withdrawal from rest to the task (i.e., parasympathetic deactivation). Among adults, ERP amplitudes elicited by to-be-attended sounds were associated with the degree of PEP shortening from rest (i.e., sympathetic activation). Overall these results suggest that greater cardiac arousal is associated with more efficient neural indices of selective attention, with higher arousal associated with more narrow attentional focus.
Talk 3: Increased neural response to wins over losses with older adults: Examining the positivity bias in aging
Anna McCarrey1,2, Joshua Goh2,3, Vijay Venkatraman4, Claudia Wolf2, Gabriela Gomez2, Susan Resnick2; 1Idaho State University, 2National Institute on Aging, 2National Taiwan University College of Medicine, 2University of Melbourne
The goal of this study was to investigate neural response to positive and negative events in a large sample of 309 healthy older adults (47.2% male; mean age: 70.3±12.2 years) from the Baltimore Longitudinal Study of Aging. Participants experienced win (positive) and loss (negative) events whilst playing a risk-taking task. For each trial, participants chose to accept or decline point offers, with a stipulated probability of winning. The instructions were to win as many points as possible. The magnitude of points offered and probability of winning varied across 72 trials. Feedback was given after each risky decision was made. Accepting an offer resulted in either a win or a loss; declining an offer resulted in neither. Two functional EPI scans were acquired per participant with 180 volumes per scan (2s TR, 37 axial slices, and 1.87*1.87*3.93 mm3 voxels). Whole brain analyses were performed using SPM8 and voxels were considered significantly activated at p < .05 corrected for the familywise error rate. Results showed significantly increased activation during feedback for wins relative to losses in several brain regions, including the visual cortex, cerebellum and posterior cingulate cortex. There were no regions that showed increased activation for losses relative to wins. Heightened neural sensitivity to wins over losses in this older sample could not be explained by age, sex or risk profile. Taken together, these findings lend support for Socioemotional Selectivity Theory and suggest a neural pathway by which older adults experience increased sensitivity for positive over negative events.
Talk 4: Attention sharpens prediction error, prediction determines behavior
Alessandro Tavano1, David Poeppel1,2; 1Max Planck Institute for Empirical Aesthetics, Frankfurt am Main, 2New York University
Attention to sensory stimuli is never uniformly distributed. In particular, attention scales up while we anticipate a stimulus that must eventually occur, improving task performance for longer awaited targets (temporal prediction). Within a predictive coding approach to perception, attention also increases the precision (signal-to-noise ratio) with which the mismatch between predicted and actual input is encoded by the brain (sensory prediction error). Here we tested whether prediction and prediction error interact in optimizing behavior or reflect distinct processes. We used the well-established auditory roving standard paradigm (Haenschel et al., 2005), in which the frequency of isochronously delivered, 50-ms pure tones changes unpredictably after a varying number of same tone standard repetitions. Perceived loudness was equalized (Impulse A-weighting). We recorded electroencephalographic data from 26 participants. The analysis focused on evoked brain responses to frequency deviants. Attention to stimuli selectively enhanced the neural encoding of small rather than large frequency differences, as reflected in the deviant N1 response. This supports the precision account and supercedes the traditional attention-capture account based on prediction error magnitude. Participants were faster in responding to longer awaited pitch changes, which also resulted in larger deviant N1 responses, highlighting the effect of temporal prediction. There was no interaction with prediction error magnitude. These results were replicated in a second experiment in which frequency was kept constant and loudness was roved. We conclude that prediction and prediction error partition concurrent but distinct influences of attention on behavior.
Talk 5: Retroactive attention can protect multiple working memory contents from perceptual interference. Evidence by event-related EEG parameters in a retro-cuing paradigm
Anna Magdalena Barth1, Edmund Wascher2, Daniel Schneider3; 1Leibniz Research Centre for Working Environment and Human Factors #1, 2, 3
To enable goal-directed behavior in changing environmental conditions it is important to keep working memory updated and to avoid distraction by irrelevant information. Focusing attention within mnemonic representations is typically studied using retroactive cues (retro-cues). So far, neither the neural correlates of protection from interference nor the amount of information that can be protected by selective attention are well understood. We addressed this issue by running EEG during a visual working memory task based on retro-cues. Participants had to memorize the angle of three differently colored bars followed by one of four retro-cue types. Two selective retro-cues indicated a subset of the memory array as being relevant for report (one or two of three bars). Additionally, two types of neutral cues were used: one cue repeated the color and position of all three bars; the other one was completely non-informative. A distractor display was presented during the retention interval in half of the experimental blocks. A distractor-induced performance decrease was only observed in neutral retro-cue trials whereas the presentation of selective retro-cues attenuated the distractor effects. Event-related potentials revealed a negative slow wave component over posterior electrodes reflecting the amount of items held in working memory after the retro-cue presentation. Moreover, a P3-like component after distractor onset in neutral retro-cue conditions indicated interference with the information held in working memory. This leads to the conclusion that selective retro-cues facilitate to an optimization of cognitive resources for preventing visual distractors from getting access to working memory.
Talk 6: Interacting long-range networks govern control over working memory
Elizabeth L. Johnson1, Callum D. Dewar1,2, Anne-Kristin Solbakk3, Tor Endestad3, Torstein R. Meling3, Robert T. Knight1; 1University of California, Berkeley, 2University of Illinois, 2University of Oslo
We investigated how frontal regions exert control over sensory mechanisms in working memory (WM). The electroencephalogram (EEG) was recorded in 14 patients with unilateral lesions localized to lateral prefrontal cortex (LPFC; age 46 ± 16) and 20 age-matched controls while they completed a visual WM task. On each trial, subjects encoded two colored shapes in specific spatiotemporal positions in preparation for a subsequent test on the identity of each shape in the pair, or on the spatial or temporal relationship between the shapes in the pair. The test prompt was presented mid-delay to initiate executive control processes (processing period). Patients exhibited impaired accuracy (87% vs. 95%, p<0.00005 ), which indicates a causal but not unitary role for LPFC in WM. Processing was marked by anterior slow (1-8 Hz) power increases and anterior-to-posterior directional connectivity alongside central-posterior alpha-beta (8-20 Hz) power decreases and posterior-to-anterior connectivity (all p<0.001). Early in processing, we observed more hemispheric asymmetry, with decreased low theta (3-4 Hz) power at the lesion site, and less anterior-to-posterior connectivity in patients (all p<0.05, corrected). We then observed that LPFC lesions could be reliably identified from parieto-occipital alpha-beta power for the remaining processing period. These results reveal an LPFC source for theta rhythms underlying executive control, a dissociable alpha-beta suppression network for WM, and a cause-and-effect relationship between LPFC theta activity and parieto-occipital alpha-beta suppression. Our findings contradict modular views of LPFC in WM, and instead demonstrate that WM is governed by the flexible recruitment of bidirectional, interacting long-range networks.
Talk 7: Prefrontal dopamine metabolism predicts neurostimulation-linked working memory training gains
Kevin Jones1,2, Jaclyn Stephens1,3, Marian Berryhill1; 1University of Nevada, Reno, 2Georgetown University Medical Center, 2Kennedy Krieger Institute
There is growing awareness that individual differences predict opposing effects on cognitive performances paired with transcranial direct current stimulation (tDCS). One possible explanation is that the effects of tDCS likely depend on individual variations in WM-relevant genetic polymorphisms, such as Catechol-O-methyltransferase (COMT val158met), Dopamine Transporter (DAT), and Brain-Derived Neurotrophic Factor (BDNF val66met). This is particularly relevant as some polymorphisms have a greater impact on cognition in older adults, a population particularly motivated to maintain WM. One hundred and thirty-seven healthy older adults provided saliva samples for genotyping and received longitudinal anodal frontoparietal tDCS (sham, 1 mA, 1.5 mA, or 2 mA) paired with 10 sessions of Visual and Spatial WM training. At baseline, significant group differences in WM performance were predicted by COMT genotype (p = .002). One month after training, there was a significant interaction of COMT genotype, tDCS intensity and WM task (p = .037). Specifically, the COMT val/val adults who received 1.5 mA, showed greater improvement on the Visual WM task, where they were initially weaker, than on the Spatial WM task, where they were initially stronger. Conversely, the COMT met/met adults in the 1.5 mA tDCS group showed the reverse pattern. Neither DAT nor BDNF were predictive of tDCS-linked WM benefits. These data suggest that intrinsic frontal dopamine activity predicts the nature of WM improvement after longitudinal tDCS. Variations in the COMT polymorphism predicted baseline WM performance and interacted with tDCS specific to WM task demands.
Talk 8: Age and Modulation of BOLD Response to Task Difficulty: the Protective Effects of Crystallized Knowledge
Zhang Jingting1, Zhuang Song1, Patricia A. Reuter-Lorenz2, Denise C. Park1; 1University of Texas at Dallas, 2University of Michigan
We have previously reported that older adults, compared to young, are less able to modulate the magnitude of the Blood Oxygenation Level Dependent (BOLD) response to task difficulty on a semantic judgment task (Kennedy et al., 2015). Here we investigated whether neural enrichment factors, such as crystallized knowledge, might be protective of modulatory capacity in the context of healthy aging. We studied 463 participants (20-89 years) who judged ambiguous (Hard) and unambiguous (Easy) words for animacy in an fMRI block design. Age and crystallized ability were entered as separate steps in hierarchical regression models to assess whether crystallized knowledge accounted for variance in modulation beyond age. We found that older adults had decreased modulation of BOLD response to difficulty in medial frontal and bilateral inferior frontal gyri (IFG), but after controlling for age, better crystallized knowledge predicted more modulation in these frontal regions. Moreover, a second analysis showed that higher modulation in these frontal regions predicted better performance in reasoning and executive function across the adult lifespan. Additionally, higher modulation in medial frontal and right IFG predicted better working memory in both the middle-aged (45-64 years) and younger-old (65-80 years) groups but not the very old group (80-89 years). These effects remained even after we corrected crystallized ability for fluid intelligence. Our findings suggested that high modulatory capability is impaired with age and is important for cognitive functions and that crystallized knowledge appears to be a type of cognitive reserve that protects modulation of brain activity in core control regions.
Talk 9: The hippocampus promotes effective saccadic information gathering in humans
Heather D. Lucas1, Melissa C. Duff2, Neal J. Cohen1; 1University of Illinois Urbana-Champaign, 2Vanderbilt University
It is well established that the hippocampus is critical for human learning and memory. Recent evidence suggests that one way in which the hippocampus contributes to learning is by allowing individuals to explore information in efficient, adaptive ways under active study conditions. Here we demonstrate that the link between the hippocampus and exploration extends even to eye movements made during an otherwise passive encoding task. In two experiments, we use the information-theoretic measure of entropy to assess the amount of randomness or disorder in participants’ item-to-item gaze transition patterns as they studied multi-item visuospatial displays. In Experiment 1, scanpath entropy at study was negatively associated with performance on a relational memory task in both within- and across-subject analyses. In particular, participants who engaged in higher-entropy viewing showed a greater tendency to later “swap” or reverse the relative positions of objects within the array. In Experiment 2, we found elevated scanpath entropy in patients with amnesia due to hippocampal damage, suggesting that the hippocampus is necessary to adaptively constrain saccadic information sampling. These data reveal that hippocampal contributions to exploratory behaviors in humans are pervasive, operating even at the level of eye movements.
Talk 10: Neural Responses Decrease While Performance Increases with Practice: A Neural Network Model
Milena Rabovsky1, Steven S. Hansen2, James L. McClelland2; 1Freie Universitaet Berlin, Germany, 2Stanford University
The observation that neural responses decrease as behavior becomes faster and more accurate with practice is ubiquitous in neuroscience. Our work provides an account of this finding in terms of a shift in the relative roles of activation and strength of connections. We used a neural network model to simulate neural responses during language understanding, and examined the model’s correlate of neural responses (specifically, the N400 component of the event-related brain potential), measured as the change in hidden layer activation induced by the current stimulus, at several time points during training of the network. We observed that the N400 magnitude first increased and then gradually decreased over the course of training while comprehension performance measured at the output layer showed a steady rise with additional practice. These results fit the empirical finding that N400 amplitudes first increase over the first few years of life and later decrease with age. Importantly, our results also speak to the issue of possible mechanisms underlying the reduction of neural activation with practice. In the model, the reduction in neural response is due to the continuous adaptation of connection weights over training. Specifically, as connection weights between the hidden and the output layer grow stronger, less activation at the hidden layer is necessary to efficiently modulate the output. This shift of labor from activation to connection weights might be an important mechanism contributing to the often observed reduction of neural activation with practice.
Talk 11: Impact of preparatory attention on subsequent memory: individual differences in cortical oscillations
Anna Khazenzon1, Shao Fang Wang1, Stephanie Zhang1, Alex Gonzalez1, Stephanie Gagnon1, Monica Thieu1, Melina Uncapher2, Anthony Wagner1; 1Stanford University, 2University of California, San Francisco
Successful encoding into episodic memory can be impaired by goal-irrelevant distractors, as well as by lapses in goal-directed attention. Individual differences in distractor filtering, both within and across subjects, may contribute to variable subsequent memory. Fluctuations in top-down goal-directed attention may have dual consequences: they may contribute to variable distractor filtering success, and may also reflect lapses of attention even in the absence of external distraction. Here, we tested these hypotheses by examining how subsequent memory for words varies with (a) the presence of external distraction (goal-irrelevant visual stimuli), and (b) EEG oscillatory measures of goal-directed preparatory attention; this neural measure provides a means of indexing attentional lapses that lead to diminished distractor filtering (when distractors are present) and diminished target stimulus encoding. High-density (128 channel) EEG was recorded during an incidental encoding task, during which participants made semantic judgments about words while ignoring infrequent peripheral images of faces or objects. A subsequent old/new recognition memory test of the target words revealed that word memory (d’) was negatively impacted by the presence of external distractors. Spectral signatures of phasic fluctuations in top-down attention – pre-stimulus posterior alpha power– predicted subsequent retrieval success, suggesting that attentional lapses contribute to memory encoding failures. Moreover, this relationship was stronger on distractor-present trials, revealing the importance of preparatory top-down attention in target stimulus encoding in the presence of distraction. These data reveal a mechanism by which individual variability in preparatory attention impacts episodic encoding.
Talk 12: Stress Effects on Memory are Context Dependent
Matthew Sazma1, Andrew McCullough1, Andy Yonelinas1; 1UC Davis
Stress after learning has been shown to improve memory for prior information in a number of studies, however there are also studies that don’t show this enhancement. In the current study, we directly manipulated both stress and context after learning to determine the significance of context for these stress effects. Results show a significant stress X context interaction, where participants in the same context showed a post-encoding stress enhancement of memory, but when context was changed between learning and stress there was an impairing effect on memory. The predominant hypothesis in the field asserts that post-learning stress enhances the consolidation processes through the release of cortisol. Analysis of the cortisol levels of participants show that even though both stress groups show similar cortisol increases, the memory effects go in opposite directions depending on the context that stress occurs in. These data suggest that current stress and consolidation theories need to be modified to account for stress needing to occur in the same context as the to-be-remembered items in order to see memory enhancements.
Talk 13: The Role of the Prefrontal Cortex in Inductive Reasoning: An fNIRS Study
Layla Unger1, Jaeah Kim1, Theodore J. Huppert2, Julia Badger3, Anna V. Fisher1; 1Carnegie Mellon University, 2University of Pittsburgh, 2University of Oxford
This study examined neural activity associated with inductive inference using functional Near Infrared Spectroscopy (fNIRS). Induction is a powerful way of generating new knowledge by generalizing known information to novel items or contexts. Two key bases for identifying targets for induction are perceptual similarity, and rules that specify category-relevant features. Similarity- and rule-based induction have been argued to represent distinct mechanisms, such that only rule-based induction requires executive function processes associated with the prefrontal cortex (PFC), namely: active maintenance of representations and inhibition of salient but irrelevant features. Here, we address the lack of direct empirical evidence supporting this possibility by recording PFC activity using fNIRS while adult participants (n=24) performed an inductive inference task. We found that PFC activity during induction was greater when participants had been taught a category-inclusion rule versus when participants could only rely on overall similarity. These results provide evidence that rule- and similarity-based induction represent qualitatively distinct processes. Specifically, rule-based induction may uniquely require executive functions associated with PFC such as the active maintenance of rules in memory, and/or inhibition of rule-irrelevant input.
Talk 14: The role of the structural connectome in literacy and numeracy development in children
Joe Bathelt1, Susan Gathercole1, Sally Butterfield1, Duncan Astle1; 1MRC Cognition & Brain Sciences Unit
Literacy and numeracy are fundamental skills acquired in childhood, a time that coincides with considerable shifts in large-scale brain organisation. However, most studies emphasise focal brain contributions to literacy and numeracy development by employing case-control designs in groups with selective deficits and voxel-by-voxel statistical comparisons. This approach is unlikely to capture the importance of broad differences in brain organisation that typically characterise brain development. In contrast, the current study was based on a more representative sample of 59 children between 6 and 16 years with varying levels of reading and maths ability, including difficulties in both domains. Further, broader differences in brain organisation were evaluated using a whole-brain structural connectome approach based on diffusion-weighted MRI data. Our results indicate an association between literacy and numeracy development in a distributed network of white matter connections that extends beyond regions implicated in a voxel-wise analysis. Further, graph theory measures of network organisation (characteristic path length, global clustering coefficient) were correlated with higher reading and maths scores. In addition, simulated disruption indicated that highly-connected regions that are particularly important for optimal network organisation also related to higher reading and maths performance. Together these findings show that changes in large-scale brain organisation contribute to literacy and numeracy development as children grow up.
Talk 15: Electrocorticography reveals the neural mechanisms of the arithmetic problem-size effect
Pedro Pinheiro-Chagas1, Amy L. Daitch2, Josef Parvizi2, Stanislas Dehaene1; 1Collège de France, Paris, 2Stanford University
Both the intraparietal sulcus (IPS) and the posterior inferior temporal gyrus (pITG) are known to be involved in arithmetic thinking, but their precise roles remain poorly understood. To characterize them, we recorded electrocorticography signals from 17 neurosurgical subjects implanted with grids of electrodes, while they were asked to verify additions such as 15+3=17. Behavioral results showed a classical problem size effect: RTs increased as a function of the size of the smallest (min) operand. We next examined how high-gamma activity is modulated by problem size. We found that the total activation across the decision period increased as a function of the min operand around the IPS, but remained constant within the pITG. Electrodes within the pITG, however, showed decreasing activation during the first second after calculation onset as a function of the min operand. Importantly, when regressing out the effect of RT, modulation of the total activity in the IPS vanished, while modulation of the initial activity in the pITG remained intact. The results suggest two distinct neural mechanisms underlying mental calculation, which would have been virtually impossible to disentangle using conventional noninvasive neuroimaging methods. While the activity in IPS seems compatible with an ‘accumulation of evidence’ pattern, similarly to neurons in the monkey LIP during a motion detection task, the activity in the pITG may represent the saliency of the evidence, similarly to neurons in the monkey MT area. Thus, we propose that the neural mechanisms previously described in basic perceptual decisions also operate during more complex symbolic reasoning.
Data Blitz Session 2
Saturday, March 25, Noon - 1:30 pm, Seacliff
Chair: Evangelia Chrysikou, University of Kansas
Speakers: Harry Farmer, Suzanne Dikker, Teodora Stoica, Arseny SOKOLOV, Andrea E. Martin, Manli Zhang, Francesca Carota, Jona Sassenhagen, Radhika Gosavi, Golijeh Golarai, Surabhi Bhutani, Andrew Quinn, Marina Bedny, Elisabeth Wenger, Brenda Rapp
Talk 1: Investigating the Neural Basis of Shared Preferences and Affiliation
Harry Farmer1, Antonia Hamilton1; 1University College London
Similarity to the self is a key factor in our judgement of others with people showing greater feelings of affiliation towards those they perceive as being more similar to themselves. We aimed to investigate the neural basis of this phenomena using an fMRI. In this study participants were required to choose which of two paintings they preferred and then observed the choices of two confederates one of whom chose the same picture as them 75% of the time while the other only chose the same 25% of the time. Behaviourally we found that participants showed greater liking to the similar confederate compared to the dissimilar confederate replicating the previous evidence for a similarity liking link. Examination of BOLD activation showed that observing the different confederate’s choice led to greater activation in both the ventrolateral prefrontal cortex, which is implicated in response switching and has previously been linked to social influence. The dissimilar confederate’s choice also led to greater activation in the dorsomedial prefrontal cortex a region heavily implicated in processing information about others. In addition we found that viewing one’s own chosen stimuli after seeing the choices of both confederates led to increased activation in regions involved in self processing and in social cognition including the temporal-parietal junction, the mid cingulate and the precuneus. Our findings suggest that the link between shared preference and affiliation involves brain areas that are involved in learning about other preferences and also regions involved in the processing of value similarity between self and others.
Talk 2: Taking hyperscanning out of the lab: Evidence from EEG recordings on 1400 dyads during face-to-face interaction
Suzanne Dikker1,2, Georgios Michalareas3, Matthias Oostrik, Hasibe Melda Kahraman4,2, Imke Kruitwagen1, Shaista Dhanesar5, Marijn Struiksma1, David Poeppel2,3; 1Utrecht University, 2New York University, 2Max Planck Institute for Empirical Aesthetics, 2Hunter College, 2Washington University in St. Louis
What does it mean to be ‘on the same wavelength’ with another person? When we feel connected or engaged, are our brains in fact ‘in sync’ in a formal, quantifiable sense? To address this question, we collected EEG and questionnaire data from 2800 participants at eight different sites (museums and galleries). During the experiment, pairs of people interacted face-to-face for 7-10 minutes inside The Mutual Wave Machine, an interactive neurofeedback art/science installation that collects, compares, and visualizes brain-to-brain synchrony between two people in real time (light patterns reflect moving-window correlations between the two EEG signals). The large dataset allowed us to explore the relationship between brain-to-brain synchrony and character/relationship traits as well as emotional states. Findings from 700 EEG recordings, matched for experimental parameters and context, show that pairs with more empathetic personalities (Interpersonal Reactivity Index, Davis 1980) also exhibited higher brain-to-brain synchrony, and the same was true for pairs who felt more connected to each other. Further, brain-to-brain synchrony increased throughout the recording session - but only if dyads were explicitly told that the light patterns they saw reflected their brain-to-brain synchrony, or if pairs reported to be more focused at the end of their session. These findings support an account whereby brain-to-brain synchrony is a possible biomarker for social interaction that increases as a function of joint attention, as measured via factors like empathy, focus, and connectedness. Our interdisciplinary ‘crowdsourcing neuroscience’ approach may provide a promising new avenue to collect rich datasets pertaining to real-life face-to-face interactions.
Talk 3: Common Neural Substrates of Down-Regulating Negative Emotion and Social Threat
Teodora Stoica1, Lindsay Knight1, Leonard Faul1, Farah Naaz1, Brendan Depue; 1University of Louisville
The relationship between the neural mechanisms underlying regulation of negative emotion and social threat perception and has not been fully elucidated. The current fMRI study aimed to investigate common mechanisms underlying the down-regulation of negative emotion and social threat perception. FMRI data was correlated with subjective behavioral ratings of negative pictures as well as fearful human faces. In addition, independent component analysis (ICA) was used to extract common networks during an emotion regulation (ER) and social threat task (STT). Behavioral results showed that ratings from the ER and STT correlated strongly. Functional results indicated a similar pattern of activation in the right hemisphere for both tasks, with robust activation in the inferior frontal junction (IFJ), medial frontal gyrus (MFG) and inferior frontal gyrus (IFG). Functional data correlated with behavioral ER ratings revealed increased rIFG and bilateral inferior parietal sulci (IPS) for better down-regulation, as well as decreased amygdala activation during the STT. Results using ICA analysis indicated common components across tasks, which loaded heavily on the right IFJ, IFG and lateral inferior parietal cortex. Selected components were also found to predict behavioral ER and STT ratings, whereby higher expression of components was related to more down regulation of negative emotion and social threat perception, respectively. The results suggest higher utilization of indicated functional networks and brain regions is related to a higher rate of down-regulating negative emotion and social threat perception.
Talk 4: The brain network for emotional body language reading: Combined structural and effective connectivity
Arseny SOKOLOV1,2, Peter ZEIDMAN2, Michael ERB3, Frank POLLICK4, Wolfgang GRODD5, Richard FRACKOWIAK1,6, Karl FRISTON2, Marina PAVLOVA3; 1Centre Hospitalier Universitaire Vaudois (CHUV), Lausanne, Switzerland, 2University College London (UCL), UK, 2University of Tübingen Medical School, Germany, 2University of Glasgow, UK, 2Max Planck Institute for Biological Cybernetics, Tübingen, Germany, 2Ecole Normale Supérieure DEC, Paris, France
While understanding the emotions conveyed by body expressions of others is essential for social cognition and interaction, the underlying neural correlates remain largely unknown. We used functional MRI (fMRI) and diffusion tensor imaging (DTI) in healthy subjects viewing a point-light arm knocking on an invisible door with different emotional expressions (happy, neutral and angry). Data pre-processing and analysis were performed with SPM12 and FSL5. The right superior temporal sulcus (STS) and right caudate nucleus exhibit higher activation for happy as compared to neutral knocking. Angry versus neutral knocking activates the inferior insula, perigenual anterior cingulate cortex (ACC) and posterior midcingulate cortex (MCC) in the left hemisphere. The cerebellar vermis (lobule IX) and right amygdala respond strongest to neutral as compared to emotional body language. To further characterize the network architecture in a neurobiologically more plausible way, we developed an approach to integrate measures of structural connectivity obtained from probabilistic tractography within dynamic causal modelling (DCM) of effective connectivity. This analysis reveals key components and interactions subserving emotional processing, such as between the caudate nucleus and cingular areas. Furthermore, the connectivity data shed light on communication between the cerebellar vermis and amygdala potentially related to emotional regulation. In summary, this study provides the first characterization of the brain network for reading of emotional body language. Combining information on function and structure is useful for network analysis and lays ground for better understanding of neuropsychiatric conditions with deficits in visual social cognition.
Talk 5: A mechanism for the cortical computation of hierarchical linguistic structure
Andrea E. Martin1,2, Leonidas A. A. Doumas1; 1University of Edinburgh, 2Max Planck Institute for Psycholinguistics
To process language, the human brain must form hierarchical representations from a sequence of perceptual inputs distributed in time. What mechanism underlies this ability? One hypothesis is the brain repurposed an available neurobiological mechanism when hierarchical linguistic representation became an efficient solution to a computational problem posed to the organism. Under such an account, a single mechanism must have the capacity to perform multiple, functionally-related computations, e.g., detect the linguistic signal and perform other cognitive functions, while, ideally, oscillating like the human brain. We show that a computational model of analogy, built for an entirely different purpose - learning relational reasoning (Doumas, LAA et al. (2008) A theory of the discovery and predication of relational concepts. Psychological Review 115:1-43.) - can parse sentences, represent their meaning, and, crucially, exhibits oscillatory activation patterns that strongly resemble the cortical signals elicited by the same stimuli (Ding, N et al. (2016) Cortical tracking of hierarchical linguistic structures in connected speech. Nature Neuroscience 19(1):158-164). Such redundancy in the cortical and machine signals suggests a deep mechanistic alignment between representational structure building and ‘cortical’ oscillations. By inductive inference, this synergy indicates that the cortical signal reflects the generation of hierarchical linguistic structure - rather than mere tracking of it - just as the machine signal does. A single mechanism – using time to encode information within a layered neural network – generates the representational hierarchy that is crucial for human language, and offers a mechanistic linking hypothesis between linguistic representation and cortical computation.
Talk 6: Language-modulated perceptual compensation: Functional connectivity analysis of L1 and L2 reading impairments in Chinese-English bilingual children
Manli Zhang1, Xiaoxia Feng2, Yue Gao2, Xiujie Yang1, Weiyi Xie1, Feng Ai1, Hehui Li2, Xingnan Zhao1, Chi Zhang1, Li Liu2, Guosheng Ding2, Xiangzhi Meng1; 1Peking University, China, 2Beijing Normal University, China
Although neural alterations associated with reading deficits in native language have attracted extensive investigation, the number of studies regarding L2 reading difficulty is still limited. In current research, we conducted an fMRI experiment wherein 74 bilingual children (i.e. 29 typically developing (TD), 23 poor Chinese readers (PCR) and 22 poor English readers (PER)) passively viewed images of Chinese characters and English words. We defined ROIs through identifying brain areas activated during both L1 and L2 processing, or selectively to either of the languages. Intriguingly, we did not observe any group differences in terms of the activation within these ROIs. Further analysis demonstrated a reduced connectivity between left inferior frontal and left supramarginal gyrus in PCR compared to TD, which was positively correlated with subjects’ phonological scores. In contrary, PCR was with a higher connectivity between left superior temporal and left superior frontal gyrus, left angular gyrus and right visual cortex, which were significantly correlated with orthographic and rapid automatized naming (RAN) performances, respectively. For PER, they showed decreased connectivity between right middle temporal gyrus (MTG) and bilateral precuneus, while their connection between bilateral MTGs was enhanced in comparison to TD, all of which were correlated with phonological performance. Together, these data suggest that children with disfluent reading tend to rely heavily on sensory cortices while processing written materials. However, this effect is subject to the modulation of different language systems. Specifically, PCR employ visual cortex to facilitate grapheme-phoneme mapping, while PER lean upon auditory processing to compensate inferior phonological awareness.
Talk 7: Representational similarity in the brain and computational language processing: New clues about the neural encoding of word meaning.
Francesca Carota1,2,3,4, Hamed Nili2,5, Nikolaus Kriegeskorte2,3, Friedemann Pulvermüller1,2,4; 1Humboldt Universit ät zu Berlin, Germany, 2MRC Cognition and Brain Sciences Unit, Cambridge, UK, 2University of Cambridge, Downing Street, Cambridge, CB2 3EB United Kingdom, 2Freie Universität, Berlin, Germany, 2University of Oxford, Oxford, UK
Language comprehension engages a distributed network of fronto-temporal, parietal and sensorimotor regions, but it is still unclear how meaning of words and their semantic relationships are represented and processed within these regions and to which degrees lexico-semantic representations differ between regions and semantic types. We used fMRI and Representational Similarity Analysis, RSA, to relate word-elicited multi-voxel patterns to latent semantic similarity among categories of action (face-, arm-, leg-related verbs) and object (animal-, food, tool-related nouns) words, as assessed by distributional statistics performed on text corpora (Latent Semantic Analysis, LSA). In left inferior frontal (BA 44-45-47), left posterior middle temporal and left precentral cortex, the similarity structure of brain response patterns conformed to the semantic similarity among action-related verbs, as well as - across lexical semantic categories - between action verbs and tool-related nouns and, to a degree, between action verbs and food nouns, but not between action verbs and animal nouns. Instead, posterior inferior temporal cortex exhibited a reverse response pattern, which reflected the semantic similarity among object-related nouns, but not action-related words. These results show that semantic similarity among categories of is encoded by a range of cortical areas, including multimodal association (e.g., anterior inferior frontal, posterior middle temporal) and modality-preferential (premotor) cortex and that the representational geometries in these regions are partly dependent on semantic type, with semantic similarity among action-related words crossing lexical-semantic category boundaries. Furthermore, these findings suggest that distributional information about word co-occurrence is relevant to shape word representations in the brain.
Talk 8: Multilayer neural network modeling of speech envelope prediction errors
Jona Sassenhagen1, Benjamin Gagl1, Christian J. Fiebach1; 1University of Frankfurt
Understanding speech critically relies on top-down, predictive processing. This allows effective encoding of bottom-up sensory information like the speech envelope. We hypothesize that the cortical auditory perception system entrains to the speech envelope in order to predict the speech envelope in the future. Then, bottom-up signaling can focus on the unpredicted sensory events (i.e. prediction errors) only. Therefore we expect that brain regions showing entrainment to the speech envelope should also reflect speech envelope prediction errors. In a naturalistic language study, we modeled the raw audio envelope of the speech stream with a 5-layer recurrent neural network. We "pre-trained" this model on multiple hours of audio books, before allowing it to "entrain" on the actual stimuli and their particular characteristics. This model described the data better than models without “pre-training” (reflecting a benefit from general knowledge of speech) or without "entrainment" (reflecting a benefit from factoring in specifics of the speaker and the acoustic environment). Prediction errors are calculated by letting the models sequentially predict the (previously ‘unseen’) speech stimulus envelope and taking the absolute of the difference to the real envelope. These prediction errors are entered into a regression model to predict continuous MEG data from 14 subjects listening to the speech stimuli, controlling for both the actual speech envelope and its first derivative. We observed that prediction errors are reliably reflected by activity at Wernicke’s area around 55 ms. This indicates that neuronal signaling of speech signals is optimized by predictive processes, allowing neuronal efficient spoken language perception.
Talk 9: A Colorful Advantage in Iconic Memory
Radhika Gosavi1, Edward Hubbard1; 1University of Wisconsin-Madison
Synesthesia is a condition in which stimulation of one sensory modality evokes experiences in a second, unstimulated modality (Simner & Hubbard, 2013). In grapheme-color synesthesia, which is experienced by 1-2% of adults, synesthetes reliably, automatically experience specific colors when viewing black-and-white graphemes. Previous case-studies have identified synesthetes with spectacular memory (Luria, 1968; Smilek et al., 2001) and group studies have found advantages for synesthetes compared to nonsynesthetes in long-term memory (Rothen et al., 2012) but have not addressed whether these advantages begin in earlier memory stages. We investigated the effect of grapheme-color synesthesia on the capacity and duration of iconic memory by testing 20 synesthetes and 20 nonsynesthetes on the Partial Report Paradigm (Sperling, 1963). We presented a letter array followed, after a variable delay, by a tone that cued participants to recall the appropriate row of the array. A repeated measures ANOVA revealed significant effects of delay (p<0.001) and group (p=0.007), but no delay*group interaction (p=0.399). Accuracy was significantly higher for the synesthetes across all delays. Furthermore, the synesthetes’ accuracy after a 500ms delay (41.4%) was almost identical to the non-synesthetes’ with no delay (42.9%)! This advantage at the earliest stage of memory implies that synesthetic experiences have perceptual underpinnings, and an enhancement in multiple memory stages. Future studies should examine the neural basis of this advantage, particularly in early visual areas, which have been shown to be involved in grapheme-color synesthesia (Hubbard et al., 2011; Gosavi et al., CNS 2016) and iconic memory processes (Sergent et al., 2011).
Talk 10: Face and place selectivity develop in tandem with the visual field representations along the VTC in children
Golijeh Golarai1, Alina Liberman1, Kalanit Grill-Spector1; 1Stanford University
In adults face-selective regions across the ventral temporal cortex (VTC) overlap with representations of central visual field, and place-selective regions with peripheral representations, perhaps due to the habitual patterns of viewing faces with central and places with peripheral vision. We previously reported the slow and differential development of category selective regions, whereby parahippocampal place selective regions (PPA) become adult-like by the teens, but face-selective regions in the fusiform gyrus (FFA) develop more slowely into adulthood. However, the developmental time course of center-periphery representations in VTC and their spatial relation to the developing category selective regions are unknown. Thus, we examined the development of center-periphery organization in the FFA and PPA in children (ages 8 - 10 , n = 12 ), adolescents (ages 12 -16, n = 13 ) and adults (ages 18 - 40, n= 12). During fMRI, subjects fixated on a central point while viewing faces and places, presented either centrally spanning 3°, or peripherally within a 12°-24° ring. Using a localizer experiment we also identified the FFA and PPA in each subject. After validating fixation performance, we found an age related increase during childhood in the peripheral bias of the left PPA and foveal bias of the FFA bilaterally, that reached adult levels during the teens. Thus, development of category selectivity and eccentricity bias overlap temporally and spatially during childhood, but face selectivity continues to develop during the teens, even after local foveal bias becomes adult like.
Talk 11: Central olfactory mechanisms underlying sleep-dependent changes in food processing
Surabhi Bhutani1, Jay A Gottfried1, Thorsten Kahnt1; 1Northwestern University Feinberg School of Medicine
Previous research suggests a strong relation between sleep deprivation (<6 h/night) and weight gain, primarily due to excessive calorie consumption from energy dense snacks. However, the neural mechanisms underlying sleep-dependent increases in appetite and food intake are currently unclear. Here we use olfactory fMRI to test the hypothesis that sleep-deprivation alters neural responses to food odors in olfactory cortex and food reward-related brain regions. In a counterbalanced crossover design, participants were randomly assigned to a night of normal sleep (8 h, NS) or a night of partial sleep deprivation (4 h, DS). A 1-week sleep stabilization phase (7- 9 h) preceded each experimental condition, which was separated by a 20-day washout period. A wrist wearable sleep monitoring actigraph verified that participants followed the assigned sleep schedules. Calorie intake was controlled on the experimental days. On the day following NS and DS, participants rated the pleasantness and intensity of individually selected savory and sweet high-caloric food odors and non-food control odors while fMRI responses were acquired. We find that across both sleep conditions, fMRI responses to food compared to non-food odors were enhanced in posterior orbitofrontal cortex, piriform cortex, and anterior insula. Critically, in the DS compared to the NS condition, fMRI activity to food vs. non-food odors was specifically enhanced in piriform and orbitofrontal cortex. By showing that early olfactory responses to food odors are elevated in a sleep-deprived state, our results highlight a role for bottom up modulation in sleep-dependent appetite and eating behavior
Talk 12: Task Evoked Dynamics in Whole Brain HMM Brain States
Andrew Quinn1, Eva Patai1,4, Diego Vidarre1,3, Anna Nobre1,2, Mark Woolrich1,3; 1Oxford Centre for Human Brain Activity, University of Oxford, 2Department of Experimental Psychology, University of Oxford, 2Oxford Centre for Functional MRI of the Brain,University of Oxford, 2Institute of Behavioural Neuroscience, University College London.
Estimation of whole brain dynamics is critical for understanding how network interactions subserve rapid cognition, yet to robustly perform such estimation requires time-series much longer than the time-scale of cognitive dynamics of interest. Here we show that Hidden Markov Model (HMM) states can characterise rapid dynamics and efficiently utilise the whole dataset to generate robust estimates of whole brain networks. This is illustrated in the context of a long-term memory paradigm involving spatial and contextual associations. MEG data were collected from 16 participants. The data were then filtered from 3-40Hz and projected into source space using a LCMV beamformer before parcellation into 44 nodes and multivariate leakage correction (Colclough et al 2015). Alpha band power envelopes were used to infer a Gaussian-HMM (Baker et al 2014;Vidaurre et al 2016) to identify transient brain states characterised by patterns of power and/or functional connectivity. Spatial maps of the relative amplitude for each HMM state were computed using the partial correlation between the state time-courses and the amplitude envelopes. Critically, the HMM decomposition is performed without any knowledge of the task conditions or timings within the dataset. To identify task-evoked changes in the HMM states, the state time-courses were epoched, and the average fractional occupancy of each state (i.e. the proportion of trials for which each state is active) was computed across participants for each point in time. The HMM approach robustly identifies dynamic brain networks across large datasets, retaining both a rich description of on-going dynamics and task-evoked responses.
Talk 13: fMRI-guided theta burst stimulation to the superior temporal cortex impairs sentence processing.
Marina Bedny1, Judy Kim1, Gabriela Cantarero2,3, Pablo Celnik2; 1Johns Hopkins University, 2Johns Hopkins School of Medicine, 2Walter Reed Army Institute of Research
Sentence processing recruits regions in inferior frontal and lateral temporal cortices (e.g., Fedorenko et al., 2011). Here, we asked whether off-line, continuous theta burst stimulation (cTBS) to these areas disrupts sentence processing. Twelve participants underwent an fMRI scan while listening to sentences and answering true/false questions. Each sentences contained two actors, and questions asked about who did what to whom. In a control condition, participants performed a working memory task with lists of nonwords. Stimulation was delivered to the left inferior frontal cortex (IFC), left superior temporal cortex (STC), or vertex. Stimulation sites were identified based on individual-subject fMRI activation during sentence processing (sentences>nonwords). We measured sentence comprehension and non-word memory performance immediately before and after cTBS. In the baseline vertex condition, performance for sentences improved following stimulation (post-pre=9%). In contrast, performance for sentences did not improve after STC stimulation (post-pre=1%) (vertex vs. STC t(11)=2.59, p<.05), suggesting that STC stimulation extinguished within-session improvement. Effects of IFC stimulation on performance did not differ from the effects of vertex stimulation. Further, the effects of cTBS to IFC, STC, and vertex on nonwords performance did not differ. Our results suggest that cTBC to STC selectively impairs sentence processing, and that cTBS may have different effects on IFC and STC.
Talk 14: Probing plasticity of auditory cortex in adulthood: Structural brain changes following pitch discrimination training
Elisabeth Wenger1, André Werner1, Simone Kühn1,2, Ulman Lindenberger1; 1Max Planck Institute for Human Development, Berlin, Germany, 2University Clinic Hamburg-Eppendorf, Hamburg, Germany
Musicians are a particularly suitable model for investigating structural plasticity of sensory processing in humans. In this study, we targeted the domain of auditory processing and investigated experience-induced changes in pitch processing. We recruited young adults between 18 and 33 years who had signed up for a course that prepares candidate students for their conservatory entrance examination. An important component of this training course is relative pitch discrimination, that is, the ability to identify tones and intervals in relation to a reference tone. Participants of the experimental group were training for different university curricula: instrumentalist (various instruments), Tonmeister, conductor, or composer (n=21). As a control group, we recruited 15 younger adults who had also received musical training in their youth and also actively performed music in their daily lives but who did not participate in a preparatory course. All participants were assessed behaviorally and with functional and structural magnetic resonance imaging (MRI) 4 to 5 times over 10-12 months. Using voxel-based morphometry (VBM) to automatically segment gray matter volume, we detected a gray matter decrease in left superior temporal gyrus in aspiring professionals compared to amateurs over time. Our results are consistent with the recently proposed expansion–renormalization model of plastic changes (Lövdén et al., 2013, NBR), and suggest that the auditory cortex of aspiring professionals who were perfecting their pitch discrimination skills was undergoing renormalization. Further analyses will focus on characterizing the shape and size of Heschl’s gyrus to asses individual variation and group differences therein.
Talk 15: Teaching cognitive neuroscience: Transformation from large lecture class to small active learning groups
Brenda Rapp1, Soojin Park1, Jeremy Purcell1, Michael Reese1; 1Johns Hopkins University
Many cognitive neuroscientists face the challenges of teaching large groups of undergraduate students the complex, rich and often difficult concepts of cognitive neuroscience. Another challenge is teaching undergraduates who enter college with extensive experience learning with digital resources and already familiar with a myriad of internet learning resources. We addressed these challenges by transforming a large (200-250 student) introductory cognitive neuroscience, lecture-style course into a hybrid format modeled on the framework developed by Steven Luck (http://psc100y.faculty.ucdavis.edu/). Course design was based on well-established learning principles such as active learning, repeated testing and distributed study. The previously twice weekly, 75-minute lectures were converted to on-line lectures that allowed for individualized viewing. Each on-line lecture was accompanied by an on-line quiz that was time-limited and deadlined. Class meetings consisted of once/week review sessions and once/week active-learning sessions; these were led by five different instructors who taught simultaneously in separate classrooms. Active Learning Sessions were the key added value of the hybrid approach allowing for experiential learning and meaningful interaction with professors. Students worked in teams on activities such as: brain navigation, lesion tracing in the language system, debating the use of CNS methods in the courtroom, building artificial neural networks and making a pitch to NIH. Course evaluation included on-line questionnaires, focus groups, classroom observations, and a comparative analysis of student responses on exam questions used in previous years. All evaluation modalities revealed significant improvement in performance and student satisfaction.