The Future of Visual Attention
30th Symposium: June 3-5, 2016
All talks & discussion sessions are in Goergen Hall
All breaks and lunches are in the Munnerlyn Atrium
Thursday, June 2
7:00 - 9:00 pm—Registration & Welcome Reception, Hilton Garden Inn
Friday, June 3
8:00 am—Registration & Breakfast
8:30 am—Welcome, Ben Hayden & Jude Mitchell, University of Rochester
Talk session I: Parietal & action
Chair: Eve De Rosa, Cornell University
Covert attention enhances the responses of neurons in many visual areas and this has widely been thought to be the neural mechanism underlying visual attention. We asked whether neuronal responses in V4 and LIP could explain the limited ability of subjects to attend multiple items in a display. We trained animals to perform a change detection task in which they had to compare two arrays of stimuli separated briefly in time. Each animal’s performance decreased as function of set-size. Neuronal responses in LIP were inversely correlated with set-size, but responses in V4 were not. In addition, neuronal discriminability in V4 was consistent across set-sizes, but was worse with higher set-sizes in LIP. The introduction of an attentional bias produced attentional enhancement in V4, but this could not explain the vast improvement in performance, whereas the enhancement in LIP responses could. These data suggest that set-size effects seen in performance and the sudden improvement in performance with focused attention may not be related to traditional attentional modulation in V4 but, instead, may be processed in LIP.
10:20 - 10:50 am—Break
Everyday life requires dynamic matching of behavior to the tasks at hand. Depending on the task, very different parts of information in the visual field should be used. This is often studied with static attentional paradigms with a single discrete response for each trial. These tasks also typically require significant training. However, this does not match well to natural task-dependent behavior, where both the goal and the distribution of information can change continuously. Thus, very little is known about how the brain accomplishes task-dependent selection and integration of information in more naturalistic continuous behavior. We therefore developed a novel paradigm which allows quantification of the temporal and spatial visual parameters that drive behavior in a continuous ocular tracking task, and which is also suitable for characterizing motion-selective neurons. Here we report behavioral results from this paradigm. The core stimulus consisted of a moving cloud of dots creating a large field optic flow (80 x 50 deg of visual angle). The focus of expansion (FOE) of the flow field moved continuously according to a random walk. To characterize the spatial integration of the visual motion, we divided the field into hexagonal subfields. Each subfield either had a small, random perturbation from the motion associated with the "true" FOE or could be blank. Both macaques and humans intuitively tracked the FOE. We used regression to determine the spatiotemporal parameters that best predicted the subject's gaze. Gaze was most influenced by the FOE location from about 250 ms before, with the majority of temporal information being integrated by 800 ms but lasting up to 2 s. Spatial integration was confined to ±4 deg near the gaze location. When another 45% of the subfields were consistent with the motion of a second, independent FOE (with differently colored dots), subjects could still track the FOE of one flow field with a similar profile of temporal integration. The paradigm provides an intuitive and easy to learn framework that can reveal the spatiotemporal profiles of motion integration during continuous behavior, and can be extended to quantitatively characterize the spatiotemporal dynamics of selection without requiring extensive training in threshold level psychophysics.
The control of visual selective attention appears to be a principal function both of parietal and prefrontal cortex, and that control involves both bottom-up and top-down modulation of sensory signals during goal-directed behavior. I will talk about recent work that addresses the relative contributions of prefrontal and parietal cortex in these two distinct forms of attention. In particular, I will focus on recent work in which we have begun testing the contribution of parietal cortical areas, including the lateral intraparietal area (LIP), to the representation of salient stimuli within prefrontal cortex, as well as on saliency-driven behavior. The results of these studies thus far suggest that parietal cortex contributes uniquely to the representation of stimulus salience.
11:50 am - 12:20 pm—Post-doc Data Blitz (3 minutes each)
12:20 - 2:00 pm—Lunch & Posters
Talk session II: Social attention and face processing
Chair: Liz Romanski, University of Rochester
A prominent feature of the primate brain is the existence of several cortical areas in which neurons respond categorically to certain visual stimuli. Most notably are the so-called "face patches", defined as regions in which fMRI responses are stronger to face images than to other types of visual objects. Recent advances in our laboratory have made it possible to record longitudinally from individual neurons across weeks and months, opening the door for new types of experiments to investigate face-selective neurons. In my talk I will describe the results of several such experiments. For example, we found that the visual selectivity of neurons in face patch AF, a "high-level" face patch, remain absolutely stable over a period up to a year, even during periods of intensive training on face identity. In agreement with several studies, I will show that neurons in this area are highly selective for faces -- in this case, based on the presentation of 10,000 flashed stimuli over a period of 3 weeks. However, the same population of "face cells", when measured during free viewing of a natural video, bear a less obvious relationship to faces and diversify significantly in their responses. Under these conditions, single neurons show a highly deterministic pattern of spiking upon repeated presentations of the same video stimulus. However, neighboring neurons are often uncorrelated in their response time courses, reflecting a different functional specialization. I will show using a novel method, combining single-unit recordings and fMRI, that individual neurons fall into several distinct categories, each bearing a unique correlative relationship with functional networks throughout the brain.
The dynamic interaction of gaze between individuals is a hallmark of social cognition. However, there is a significant knowledge gap in our understanding of social attention with respect to contingent gaze dynamics. Moreover, the potential role of neuromodulation in social interaction remains unclear. We utilized a live gaze interaction paradigm in pairs of rhesus macaques to study contingent gaze dynamics between two individuals and modeled these dynamics as a decay process of sustained attention following mutual eye contact. Furthermore, we examined how the oxytocinergic and opioidergic systems interact to modulate these contingent gaze dynamics. Dominance and familiarity between the pairs induced separable components of gaze dynamics. Following mutual gaze, dominance promoted an enhanced level of prolonged social attention, whereas familiarity promoted a higher initial decay rate. Following inhaled administration of either oxytocin or naloxone, an opioid antagonist, or both, we found that oxytocin co-administered with naloxone more strongly promotes attention to the eyes of a conspecific’s face compared to either drug administered alone. Preliminary analyses revealed that the effects of oxytocin and naloxone were simply additive for non-contingent looking behaviors, such that the added effects of oxytocin and naloxone alone mirrored the effect of oxytocin and naloxone administered together in a one-to-one positive linear correlation. As opposed to the linear additivity observed for non-contingent gaze, the combination of oxytocin and naloxone invoked a supralinear enhancement of prolonged social attention following mutual eye contact compared to oxytocin or naloxone alone, such that administering the two agents together produced a significantly larger effect than the summation of the effects observed when the drugs were administered separately. Our findings are supported by the known regulatory relationship between the oxytocin and opioid systems, in which attenuated opioid processing is associated with stronger oxytocin release from the posterior pituitary. We provide the first preliminary evidence that the oxytocin and opioid systems interact to modulate social attention and exploration following contingent social interactions based on the observed supralinearly summed effect patterns.
The primate amygdala plays a key role in processing visual stimuli with social significance, but it remains debated whether this depends on attention. During natural vision, fixations indicate the location of attention and amygdala neurons are expected to respond to the social significance of fixation targets. We recorded from amygdala neurons in both humans and macaques during free viewing of arrays of images. Amygdala neurons responded transiently only during fixations on preferred stimuli in both species, with more selectivity in humans. The majority of neurons preferred faces of conspecifics, with fewer responding to heterospecific faces -- a pattern mirroring first fixation preferences. Response latencies were shortest for conspecific-preferring neurons, indicating that the amygdala selects for processing details with species-specific social significance. This comparative work demonstrates for the first time that primate amygdala neurons have small effective receptive fields when viewing multiple stimuli that compete for attention.
3:30 - 4:00 pm—Break
Talk session III: Reward and learning
Chair: Kari Hoffman, York University
Self-control depends on the ability to prioritize different considerations in different contexts. For example, losing weight requires greater focus on a food’s healthiness and less on its tastiness. Acting generously requires attending to social considerations and ignoring egoistic concerns. The neurocomputational basis for attentional modulation of value-based choice remain poorly understood, and it is unclear whether it is supported by domain-general or domain-specific mechanisms across choice domains. We combined computational modeling with multivariate decoding of fMRI data to determine how the value of choice attributes in different choice domains are reweighted under different attentional goals. Participants (n = 36) completed two choice tasks on separate days involving either foods varying in healthiness and tastiness or monetary proposals varying in benefits for self and others. In both tasks, participants made choices under “natural” and “attentional focus” conditions (e.g., “respond naturally”, “focus on the food’s healthiness”, “focus on your partner’s feelings”). Behaviorally and neurally, we observed evidence for both domain specificity and domain generality. Supporting domain specificity, the ability to focus on others’ welfare when instructed to do so amplified neural coding of that attribute in a distinct set of social cognitive brain regions, and was uncorrelated with success in regulating other choice attributes, either within or across domains. Supporting domain generality, the ability to successfully suppress choice attributes across domains (e.g., tastiness, selfish considerations) was highly correlated at the behavioral level, and represented neurally via flexible coding of these attributes in a single, overlapping region of the dorsolateral prefrontal cortex that has previously been implicated in value-based choice. These results point toward exciting new questions about when and why attentional modulation of decision-making operates via different mechanisms.
The talk will discuss approaches aimed at understanding the computational mechanisms that drive learning and development in young children. Although infants are born knowing little about the world, they possess remarkable learning mechanisms that eventually create sophisticated systems of knowledge. We discuss recent empirical findings about learners’ cognitive mechanisms—including attention, curiosity, and metacognition—that permit such striking learning throughout infancy and childhood. We will review evidence that infants enter the world equipped with sophisticated attentional strategies that select intermediately complex material to maximize their learning potential (the “Goldilocks effect” of infant attention, e.g., Kidd, Piantadosi, & Aslin, 2012, 2014; Piantadosi, Kidd, & Aslin, 2014). We will also discuss more recent work on the dynamics of idealized attention in complex learning environments, with a focus on attentional-switching patterns and their implications for understanding learning (e.g., Pelz, Piantadosi, & Kidd, 2015; Pelz, Yung, & Kidd, 2015). Finally, I discuss the importance of understanding these basic cognitive mechanisms in order to eventually develop more efficient educational frameworks, better understand learning delays, and discover the causes and best interventions for cognitive disorders like ADHD.
Economic choice imposes an important binding problem on mental and neural systems that implement it. This binding problem comes in three forms, the need to bind values to options, to bind chosen options to corresponding actions, and to bind outcomes with choices in post-reward learning. Standard approaches often assume that the reward system uses a labelled line approach to solving these binding problems. We explore the alternative possibility, that the brain binds options, values, and outcomes using selective attention. This possibility makes several counterintuitive predictions, most notably that we don't decide between options so much as we make a series of accept-reject decisions for each option in sequence.
6:00 - 9:00 pm—Banquet, Hilton Garden Inn
Saturday, June 4
8:00 - 9:00 am—Breakfast
Talk session IV: Circuits and gating (Part I)
Chair: Lorella Battelli, Harvard Medical School/IIT, Italy
Attention allows us to select relevant sensory information for preferential processing. I will discuss effects of attention on early visual processes. I will present psychophysical and fMRI studies regarding the effects of endogenous (voluntary) and exogenous (involuntary) covert attention –the selective processing of visual information without eye movements– on the perception of basic visual dimensions. Specifically, I will showing how contrast sensitivity increases at the attended location (or for the attended features) at the expense of reduced sensitivity at unattended locations (or for the unattended features), and discuss these results in reference to a normalization model of attention.
Attention is not a singular phenomenon but a complex construct for which operational definitions and common usage can vary enormously. Are states such as quiet alertness, vigilance, and focal attention best viewed as existing on a spectrum (and thus differ in degree but not type)? Or as being distinct processes with distinct neural mechanisms? There is good evidence to support a role for the neuromodulator acetylcholine in the neural processes supporting vigilance; that is attentiveness as a state rather than selective attention to a particular object or region of space. Not much is known, however, about how acetylcholine serves this function, at a circuit level. Also not known is which attentive behaviors can be accounted for (in part or entirely) by cholinergic modulation. We study cholinergic modulation of cortical circuits with the goal of addressing both these unknowns, using the visual cortex of the macaque monkey as our model system. The choice of model system is crucial in this case, as species differences in the anatomy of the cholinergic system are profound. Our studies of anatomy and in vivo physiology and pharmacology have shown that acetylcholine - acting via pre-synaptic nicotinic receptors at the thalamocortical synapse in layer 4c of area V1 - increases the gain of the ascending sensory drive to the cortex. The result is a multiplicative response gain reminiscent of that seen in some behavioral studies of focal visual attention. This modulation at the entry point to the cortical circuit appears to be conserved across sensory systems and species. Activating muscarinic receptors, on the other hand, yields a mixed enhancement and suppression of cortical circuits that differs across cells within an area, between cortical areas, and across species. These differences in receptor expression across cortex yield distinct modulatory compartments, between which signaling conditions can differ substantially. In addition to producing diverse responses in the receiving circuit, this variation in receptor expression suggests mechanisms that locally modify the release and diffusion of acetylcholine. In order to better understand the differences between these signaling compartments, and the behavioral drivers of acetylcholine release, we are developing tools that enable the local sensing of acetylcholine concentration on fine timescales in behaving macaques. The deployment of these tools will open up new ways of understanding and exploring the dynamic control of cortical circuits by acetylcholine, and other diffuse signaling molecules.
10:00 - 10:30 am—Break
Neurons throughout the visual system are not uniformly modulated by attention. There are examples of neurons in all visual cortical areas whose firing rates are increased when attention is directed toward stimuli within their receptive fields and there are also many examples of neurons whose firing rates are unchanged or are even decreased by attention directed toward the receptive field. I suggest that we can learn about the rules governing attentional modulation – and the underlying neural mechanisms of attention – by examining diverse effects of attention on neuronal activity in the visual cortex. My approach is to study how visual spatial attention modulates the activity of neurons in the primary visual cortex (V1), the first cortical area to process visual information. While attentional modulation of V1 neurons is weak on average, there is wide variation in attention effects on V1 neurons, including facilitation and suppression of neuronal firing rates. Furthermore, given the vast amount of information available about the local circuitry and visual physiology of V1 neurons, it is possible to ask whether attentional modulation of V1 neurons depends upon neuronal feature selectivity and/or position within the V1 cortical architecture. We find a number of striking correlations between laminar position, feature selectivity, and attentional modulation within V1. Overall, our results suggest that attentional modulation depends upon the match between feature selectivity of V1 neurons and feature detection required for the task. We also observe systematic relationships between attentional modulation and position of V1 neurons within the local cortical architecture. Together, our results suggest that in order to understand the neural mechanisms of attention, we must examine attentional modulation at the granular and circuit levels, taking into account neuronal feature selectivity and task demands.
Goal-directed attention allows for the prioritization of behaviorally-relevant sensory information and has been shown to improve both perceptual sensitivity and encoding of attended stimuli in sensory cortex. These attention-dependent changes in encoding differ according to the relationship of the neurons’ tuning to the goals of the task, suggesting that perceptual changes are due to selective changes in the activity of functional subgroups of sensory cortical neurons. To determine if attention modulates specific subnetworks we monitored the activity of the diverse population of neurons in the primary visual cortex (V1) of mice performing a modality-specific, goal-directed attention task. In this task, head-fixed mice were cued on a trial-by-trial basis to switch between detecting changes in the orientation of a visual stimulus or the volume of an auditory stimulus; notably, visual stimuli before the change were symmetrical across trial types. We then tested the perceptual effect of attention by presenting rare, invalidly cued trials in which the mice were rewarded if they detected a change in the uncued modality. Mice had a lower probability of detecting invalidly cued targets, demonstrating that the cue is sufficient to set the expectation of, and attention toward, changes in the cued modality. Additionally, mice were less likely to respond to invalid targets on longer trials, indicating that they maintain, or even increase, their goal-specific attention as the trial progresses. In order to investigate the rapid neuronal modulation that occurs during task performance we monitored neuronal activity in V1 using two-photon imaging of GCaMP6. We analyzed two distinct epochs during task performance: the responses to the baseline stimuli (during anticipation of the stimulus change) and to the target stimulus. We found that, on visual trials the average responses to the baseline stimuli were transiently suppressed while responses to the target were enhanced. Interestingly, both of these effects were driven more by neurons tuned to the target stimuli than to the baseline stimulus. This suggests that attentional modulation in V1 is specific to neurons that provide the most information about the visual change. Investigation into the local and long-range circuits these neurons belong to will further elucidate the relationship of primary sensory processing to perception and behavior.
11:30 am - 12:00 pm—Grad Student Data Blitz (3 minutes each)
12:00 - 2:00 pm—Lunch & Posters
Talk session V: Circuits and gating (Part II)
Chair: Ralf Haefner, University of Rochester
Spatial selective attention is known affects response baseline, response gain and correlational statistics of neurons in the early and intermediate visual systems. These modest effect (~15% in area V4) likely serve to increase the SNR of stimuli at attended spatial locations. However, as information is pooled across successive layers of the cortical hierarchy, these small changes in response baseline and gain must inevitably lead to changes in tuning of single neurons at relatively later stages of processing. Indeed, feature-based attention affects tuning even in area V4, and there is ample evidence for task-related changes in tuning in prefrontal cortex. The tuning changes observed in single neurons imply that attention actually changes the way that signals are represented throughout the cerebral cortex, expanding the cortical representation of attended stimuli at the cost of compressing the representation of unattended stimuli. In other words, it is likely that attention does not merely increase SNR, but it changes the very way that information is represented throughout the cortical hierarchy.
Finding sought visual targets requires our brains to flexibly combine working memory information about what we are looking for with visual information about what we are looking at. To investigate this comparison process, we recorded neural responses in inferotemporal cortex (IT) and perirhinal cortex (PRH) as macaque monkeys perform tasks that require them to find visual targets in sequences of distractors. Our results reveal the existence of computations in PRH that act on input arriving from IT to produce a signal that indicates when a target is found. Additionally, input-output models fit to our data reveal that these PRH computations can be accounted for by a relatively simple process. These results suggest that much of the complexity reflected in the PRH responses is thus likely inherited from PRH inputs.
Decision-making is a complex process in which different sources of information are combined into a decision variable (DV) that guides action. Neurophysiological studies have typically sought insight into the dynamics of the decision-making process and its neural mechanisms through statistical analysis of large numbers of trials from sequentially recorded single neurons or small groups of neurons. However, detecting and analyzing the DV on individual trials has been challenging. Here we show that by recording simultaneously from hundreds of units in pre-arcuate gyrus of macaque monkeys performing a direction discrimination task, we can predict the monkey’s choices with high accuracy and decode DV dynamically as the decision unfolds on individual trials. This advance enabled us to study changes-of-mind (CoM) that occasionally happen before the final commitment to a decision. On individual trials, the decoded DV varied significantly over time and occasionally changed its sign, identifying a potential CoM. Interrogating the system by random stopping of the decision-making process during the delay period after stimulus presentation confirmed the validity of identified CoM. Importantly, the properties of the candidate CoM also conformed to expectations based on prior theoretical and behavioral studies: they were more likely to go from an incorrect to a correct choice; they were more likely for weak and intermediate stimuli than for strong stimuli; and they were more likely earlier in the trial. We suggest that simultaneous recording of large neural populations provides a good estimate of DV and explains idiosyncratic aspects of the decision-making process that were inaccessible before.
3:30 - 4:00 pm—Break
To successfully interact with the environment the brain must select and filter out behavioral relevance information from the irrelevant, a process known as selective attention. In primates, the lateral prefrontal cortex (LPFC), located anterior to the arcuate sulcus and posterior to the principal, it is thought to play a fundamental role in the voluntary allocation of attention. We recorded the responses of many neurons in this area using multielectrode arrays while macaque monkeys allocated attention to a target stimulus in the presence of distracters. A large proportion of neurons were selective for the allocation of spatial attention to one location in the visual field. We show that the allocation of attention across the visual field could be accurately estimated, on a single trial basis, from the simultaneous activity of LPFC neuronal ensembles using a linear decoder. The decoding accuracy was substantially larger than the one provided by single neurons, and it was strongly influenced by the size and composition of neuronal ensembles. Finally, the correlation structure of the neuronal ensembles influenced decoding accuracy also depending on the ensemble size and composition.
The selection of information from our cluttered sensory environments is one of the most fundamental cognitive operations performed by the primate brain. In the visual domain, the selection process is thought to be mediated by a static spatial mechanism – a ‘spotlight’ that can be flexibly shifted around the visual scene. This spatial search mechanism has been associated with a large-scale network that consists of multiple nodes distributed across all major cortical lobes and includes also subcortical regions. To identify the specific functions of each network node and their functional interactions is a major goal for the field of cognitive neuroscience. In my talk, I will show behavioral and neural evidence that the attentional spotlight is neither stationary nor unitary. In the appropriate behavioral context, even when spatial attention is sustained at a given location, additional spatial mechanisms operate flexibly in parallel to monitor the visual environment. These rhythmic processes can be linked to a specific fronto-parieto-thalamic network. This evidence indicates the need for major revisions of traditional attention accounts.
5:30 - 8:30 pm—Poster Session/Grazing Dinner, Munnerlyn Atrium
Sunday, June 5
8:00 - 9:00 am—Breakfast
Talk session VI: Executive control
Chair: Tania Pasternak, University of Rochester
Cognitive control refers to the allocation of mental resources to optimize the performance of goal-driven behavior. Many of our decisions in daily life require the engagement of cognitive control processes. For example, when approaching a traffic intersection in which the signal turns yellow, cognitive control processes must engage in order to make the optimal decision of hitting the brake, hitting the accelerator, or cruising. Examples of these processes include maintaining the context of long-term goals in working memory (Am I in a rush? How many points are on my driving license?), focusing attention on relevant environmental parameters (speed, distance to intersection, presence of police vehicle or traffic camera) while ignoring irrelevant ones (kids yelling in the back seat), choosing responses concordant with long-term goals (avoiding large fines) while overriding prepotent or automatic responses (get there as fast as I can), monitoring the outcome of previous behavior, and adjusting future behavior to maximize profit and minimize pain.
The prefrontal cortex (PFC) is known to play an important role in controlled decision-making. We have studied these processes in the human brain using the opportunities afforded by neurosurgical procedures requiring intracranial neurophysiological recordings. Patients undergoing either deep brain stimulation for movement disorders or intracranial electrode implantation for seizure monitoring consented to these studies with institutional IRB approval. They performed the multi-source interference task (MSIT), a Stroop-like task requiring controlled decision-making.
We find that individual neurons in the dorsal anterior cingulate cortex (dACC), a region within the medial PFC (mPFC), encode key aspects of these control processes. Approximately one-quarter of the ~200 recorded dACC neurons encode task difficulty or “conflict”, a parameter that would be required in order to appropriately allocate control. Another approximately one-quarter of the dACC neurons encode response selectivity, preferentially changing their firing rate in accordance with the correct response. Approximately one-fifth of the neurons are sensitive to the presence of feedback, modulating their firing rate in a pattern consistent with an unsigned prediction error. We also recorded local field potentials (LFPs) from dACC and dorsolateral PFC (dlPFC). Granger causality and conditional mutual information analyses revealed that feedback information is transferred from dACC to dlPFC, preferentially so by modulation of oscillatory power in the theta band. Finally, we recorded spiking activity in the dlPFC and find that as opposed to dACC neurons, these neurons preferentially encode information about task difficulty not in firing rate, but rather in spike-theta coherence.
Taken together, these findings suggest that key aspects of controlled decision-making, including monitoring of task difficulty and feedback, are encoded in firing rate changes in dACC neurons, and that this information is conveyed to dlPFC using theta synchrony. These findings have implications for refining models of cognitive control processes in human PFC.
Neurons in primate prefrontal cortex control the formation of large-scale attention networks, but how the activity of cells and circuits exerts network control is unknown. This talk outlines possible mechanisms on how rhythmically structured activation of single neurons links to large scale network dynamics.
I will first provide empirical evidence including an illustration that burst-synchronization serves as potent candidate mechanism to achieve network control during attentional states. Prefrontal burst firing of single neurons increases during attentional states, synchronizes long-range those prefrontal and anterior cingulate cortices that also functionally interact during controlled behavior, and is subserved by specific cell-types - with interneuron bursts likely triggering remote LFP coherence at beta frequencies and pyramidal cell bursts reflecting most likely coherent dendritic-input.
This example-case of attention-specific burst synchronization gives rise to a Dynamic Circuit Motifs hypothesis that predicts a unique linkage of (1) the (burst-) synchronized activation state, (2) the underlying structural components (cell types and dendritic currents), and (3) the generic computational function implemented by the activated motif (context-dependent gating). The talk discuss the challenges and merits of a Dynamic Circuit Motifs perspective to advance our understanding of long-range attention networks.
10:00 - 10:30 am—Break
Under a widely-accepted framework in sensory neuroscience, perceptual decisions are limited by noise in sensory encoding. However, a long tradition in cognitive psychology emphasises that decisions are limited by the capacity of information processing. I will describe a computational framework that unifies these two views, with a focus on how cues providing advance information about the probability of sensory signals (expectation) and their relevance (attention) modulate perceptual decisions.
Computational theories propose that spatial attention modulates the topographical landscape of ‘priority’ maps in regions of visual cortex so that the location of an important object is associated with higher activation levels. Consistent with this framework, single-unit recording studies reveal attention-related increases in the gain of neural responses as well as changes in the size of spatial receptive fields. However, the joint impact of these modulations on the overall fidelity of region-level priority maps is not well understood. I will speak about our work using fMRI to reconstruct spatial representations of attended and ignored stimuli using large-scale activation patterns measured from areas of occipital, parietal and frontal cortex.
12:30 - 2:30 pm—Closing Lunch at the home of Jude Mitchell