Robert T. Knight - US grants

Extensive research supports a dominant role of mammalian prefrontal cortex (PFCx) in the selection of sensory stimuli and in control of goal directed behavior. Behavioral and physiological studies reveal an inhibitory PFCx output to diverse neural systems with PFCx regions functioning to both filter irrelevant sensory inputs and to inhibit inappropriate behavior. In non-human primates with PFCx ablation, distractibility due to intrusion of irrelevant neural inputs produces an inability to maintain focused attention with subsequent deterioration of behavioral performance. Although observation of patients supports a similar role of human PFCx in the control of attention allocation, little information on the underlying physiological mechanisms is available. It is proposed that, as in animals, inability to control early stages of sensory processing coupled with abnormalities in the maintenence of attention and orientation systems are the basis of the disorganized behavior apparent after human PFCx damage. Recent studies of auditory event-related potentials (ERPs) from our laboratory confirm abnormalities in both attention and orientation capacity in humans with PFCx damage. Experiments are proposed to further define these deficits and to study the parallel phenomena in the somatosensory and visual modalities. If PFCx is a critical supramodal region, lesions of this area should produce deficits in the selection of sensory information associated with attention and orientation abnormalities irrespective of the sensory modality. To address this issue, experiments on PFCx lesioned humans are proposed in the auditory, visual and somatosensory modalities to test three interrelated hypothesis: 1) is there a deficit in the control of sensory inputs as manifested by abnormal increases in amplitude of sensory evoked potentials?, 2) are there hemispheric aysmmetries in attention mechanisms as revealed by decrements in the negative electrical shift generated to selectively attended inputs?, and 3) is there evidence of a decreased orienting response to novel inputs as indexed by a decrement in the N200-P300 complex generated to unexpected and deviant stimuli? These studies may provide insight into the contribution of human PFCx to the control of stimulus selection, orientation and attention systems employing non-invasive techniques and may permit more objective evaluation of prefrontal function in neurological and psychiatric populations.

A P300 event-related potential, the so-called "decision wave", is generated when attention is drawn to infrequent environmental events. The P300 is widely used in clinical and basic psychological research to draw inferences about human cognition. Therefore, the neural origin of the P300 is of considerable interest to cognitive neuroscientists and clinicians. Studies in our laboratory have shown that unilateral lesions in prefrontal or temporal-parietal association cortex can reduce or abolish components of the auditory P300. A variety of evidence, including intracranial recordings from hippocampal formation, suggests an important role of limbic structures in P300 generation. Neuroanatomical studies in primates have documented reciprocal connections through which prefrontal and temporal-parietal cortex could modulate hippocampal function. The scalp P300 may index the engagement of this association cortex-hippocampal circuit, which is activated during the detection and encoding of biologically significant stimuli. A series of studies is proposed to investigate the contributions of association cortex and hippocampus to P300 generation. Groups of patients with focal infarctions of prefrontal or posterior association cortex, or infarctions of the posterior hippocampus will be studied. Parallel detection experiments with graded levels of difficulty will be performed in auditory, visual and somatosensory modalities. High resolution MRI neuroimaging will be used to reconstruct lesion extent in all patients, and to evaluate possible modality-specific elements of P300 generator circuits. Concurrent analyses of reaction times and response accuracy will be used to characterize the cognitive concomitants of the P300 abnormalities. The proposed research plan is designed to provide systematic data on the neural and behavioral substrates of the P300.

Event-related potentials (ERPs) provide millisecond temporal resolution of neural activity underlying cognitive processing. Research utilizing implanted electrodes in behaving humans, animal models, neural modeling of intracranial sources and study of neurological patients with focal damage in neocortical or limbic regions has provided converging information on the intracranial generators of ERPs. Thus, ERPs constitute a physiological method for measurement of cognitive activity with high temporal and spatial resolution. Study of patients with focal brain damage has shown that the P300 response, initially proposed to be a unitary phenomenon, has multiple modality independent generators in cortex in prefrontal regions, temporal-parietal junction and hippocampal formation. Patient research utilizing combined behavioral and electrophysiological techniques may provide additional information on the cognitive basis of different ERP components since specific ERP amplitude reductions can be lined to defined behavioral deficits. Theories of the cognitive basis of P300 have centered around attention and memory although no consensus has emerged. The N400 response has been linked to linguistic processing and access to the long-term memory store. Less information on neural sources is available although intracranial recording has reported N400-like activity in entorhinal cortex. Recent data collected in our laboratory in neurological patients with focal areas of brain damage suggest that the P300 and N400 responses may index neural activity generated in neocortical and limbic regions during early attention and working memory processes. In the current proposal we will study groups of patients with MRI defined damage in subregions of either prefrontal or posterior association cortex or in mesial temporal areas including hippocampus and adjacent entorhinal cortex. Combined ERP and behavioral methods will be employed to test cognitive theories of P300 and N400 generation. This approach may also provide further insight into the neural regions responsible for generation of the P300 and N400.

DESCRIPTION:(from applicant's abstract) This proposal is designed to further our understandingof thecontribution of prefrontal cortex (PFCx) to human cognition. Neuropsychological and electrophysiological techniques will be employed to examine the role of subregions of PFCx in executive control of visual processing, novelty detection and response monitoring. 1) Experiments will be conducted to delineate the temporal (100-600 rnsec) and spatial dynamics of lateral PFCx-extrastriate interactions during visual attention using event-related potentials (ERPs) and EEG spectral techniques. 2) We will employ EEG spectral techniques to examine the temporal parameters of lateral PFCx-extrastriate interactions during the switching of attention to subregions of the visual field. 3) There is disagreement as to how object andspatial information is integrated in lateral PFCx. Behavioral andelectrophysiological experiments will be performed to assess the contribution of lateral PFCx to the integration of visual features. 4) The contribution of lateral PFCx andorbital PFCx to object, spatial and emotional working memory will be explored. 5)A distributed cortico-limbic network involving PFCx, anterior cingulate, temporal-parietal junction and posterior hippocampal formation is activated during novelty detection (200-500 msec). This network provides a neural measure of phasic attention to perturbations in the environment essential for mental flexibility andnew learning. Behavioral end electrophysiological experiments will examine how lateral and orbital PFCx contribute to different aspects of novelty processing. Experiments will examine the contribution of emotional valence as well as context to novelty processing. 6) Finally, behavioral performance is constantly monitored and detection of mistakes leads to reliable alterations in current andfuture action. In humans, detection of an error is accompanied by an error related potential (ERN) generated in anterior cingulate cortex within 100 msec after detection the erroneous response. We have shown that this cingulate measure of response monitoring is regulated by lateral PFCx. Two theories including response competition and the emotional valence of a detected error have been proposed to explain the role of prefrontal-cingulate circuits in response monitoring. Experiments will be conducted in patients with lateral or orbital PFCx damage to test these hypothesis. This proposal derives from ongoing research in our laboratory designed to elucidate the temporal dynamics and network properties of PFCx contributions to cognitive processes frequently impaired in neurological and psychiatric disease.

With National Science Foundation support, Ms. Beer and her advisor Dr. Knight will conduct a year-long investigation of brain areas that may be involved in successful social interaction. Human social interaction occurs in specific contexts in which there are norms dictating which behaviors are appropriate. Previous research suggests that the orbital prefrontal cortex may be involved when individuals restrain their social behavior. This area is thought to be active when individuals inhibit their emotional behavior or use emotional information to make social judgments. In two studies, Beer and Knight will examine whether orbital prefrontal cortex mediates emotional regulation (i.e., the ability to control the expression of emotion), whether the regulatory processes of orbital prefrontal cortex extend to self-disclosure (i.e., any information about oneself that is personally communicated to another individual), and whether the orbital prefrontal cortex is important for utilizing emotional information when making social judgments? Beer and Knight will compare individuals with orbital prefrontal cortex damage to individuals with intact orbital prefrontal cortex. Both groups will be asked a series of personal questions by a stranger and subsequently watch a videotape of this interaction. If the orbital prefrontal cortex is involved in restraining social behavior in accordance with social norms, the group with orbital prefrontal cortex damage will inappropriately express emotion and disclose more personal information than the healthy group. Similarly, if the orbital prefrontal cortex is involved in applying emotion information to social judgments, the group with orbital prefrontal cortex damage will inaccurately judge their behavior after watching the videotape of their social interaction.

This research is likely to increase knowledge about the brain basis of emotional regulation, self-disclosure, and the application of emotional information to behavioral judgments. These behavioral and brain functions have significant roles in personal and social life.

DESCRIPTION (provided by applicant): The Program Project employs a multidisciplinary approach focused on delineating the neural mechanisms supporting cognition in humans. Three key components of human cognition including executive control, memory and language will be studied. These three Projects interact extensively in an effort to bridge theoretical and experimental boundaries between research domains. The principle goals of the Program Project are to define discrete processes unique to particular cognitive operations and to elucidate common neural mechanisms enabling humans to fluidly perform a range of cognitive operations. To achieve this, we will utilize several powerful research methodologies in each project. The central component of the Program Project is an extensive neurological patient population with well-defined focal brain damage. This research population has been developed and maintained over the last 21 years at UC Berkeley, UC Davis and the VA Northern California Health Care System and forms a unique neuropsychological tool for studying human cognition. The Core of the Program Project provides detailed neuropsychological and neuroanatomical definition for all the neurological patients to be investigated in the specific projects. The Core also supports and extends state of the art human electrophysiological and functional brain imaging facilities available to all researchers in the Program Project. Project 1's focus is on executive control of cognition with a particular emphasis on the role of sub regions of prefrontal cortex in attention, working memory and task switching. The role of prefrontal cortex in executive control of memory and language is also examined. Project 2 addresses the role of medial temporal regions in memory storage and binding of new information. A prefrontal-medial temporal system has been implicated in the preferential detection and long-term storage of novel information. These processes will be explored in a series of experiments aimed at understanding the neuroanatomical and temporal relations of cortico-limbic interactions in memory processing. Project 3 addresses key questions on the neural organization and functional interplay of distributed cortical networks engaged in language processing. This project draws on an extensive aphasic population in combination with electrophysiological and fMR1 techniques in an effort to study key linguistic issues related to articulation, lexical-semantic storage and access, and executive control of language. The Program Project fuses cognitive neuroscientists with the behavioral and physiological tools critical for definition of the temporal and neuroanatomical substrates of the cognitive processes central to normal and disordered human cognition.

DESCRIPTION (provided by applicant): The aim of this Program Project is to employ state of the art behavioral, neuroanatomical, and electrophysiological and hemodynamic methodologies to study the neural basis of cognition. Converging experimental approaches will enhance our understanding of the spatio-temporal properties of the cognitive processes under investigation. The Core initially describes the administrative structure and key personnel and then provides an overview of the research environment. The Core then delineates the research methodologies and infrastructure common to the Executive Control, Memory and Language Projects. A detailed description of the patients is provided as well as the methods used for selection and definition of the neurological and behavioral status of all research subjects. Neuroanatomical methods for lesion definition including extent of cortical damage and new techniques for mapping white matter tract deterioration are provided. Finally, physiological techniques to be employed including EEG/ERP and fMRI procedures are comprehensively reviewed. The Core will also support exploration of novel methods for neuroanatomical lesion reconstruction, fusion of EEG and fMRI data. The Core will also delineate parameters that will need to be considered in the study of neurological patients with fMRI approaches. The infrastructure provided in the Core will provide the advanced neurobiological tools necessary to study human cognition in normal and neurologically impaired populations.

DESCRIPTION (provided by applicant): This proposal is designed to further our understanding of the contribution of the orbital prefrontal cortex (PFCx) to human social behavior. Techniques drawn from cognitive neuroscience will be used in an effort to integrate information processing theories of social adjustment with theories of emotional influences on cognition. Neuropsychological and high field functional neuroimaging (4T) approaches will be employed to examine the contribution of orbital PFCx to the interplay of emotional response on two on-line predictors of social adjustment: decision making and judgment of social stimuli. Neuropsychological studies in patients with discrete damage in orbital PFCx examine the involvement of this region in emotional biasing of social decisions and judgments. Functional neuroimaging studies in healthy controls will assess whether there is evidence of activation in orbital PFCx in these tasks. The experimental approach is designed to provide converging lesion and fMRI information on the role of orbital PFCx in social behavior. The proposed research draws on paradigms for eliciting emotion and examining decisions and judgments that have been used extensively in previous research. Well-validated tasks assessing risky decisions and judgments of social behavior that are susceptible to emotional bias will also be implemented. Risky decision making will be assessed using the Perceptions of Risk Questionnaire and a roulette gambling task. Judgments of social behavior will be assessed using the Profile of Nonverbal Sensitivity and the Pictures of Facial Affect Set. The proposed research represents a novel application of these paradigms because previous studies have not attempted to systematically relate these processes to activity in brain areas felt to be implicated in social adjustment. The proposal is designed to provide a basis for developing a larger program of research aimed at understanding the neural basis of social behavior.

[unreadable] DESCRIPTION (provided by applicant): Neurosurgical patients who need a brain tumor, epileptic region or other structural lesion removed that might be near language cortex require an invasive 45 min procedure (intra-operative electrical stimulation mapping; ESM) to map language and motor function and guide the neurosurgeon's operative approach. The information gained from ESM is critical to spare language areas and has been shown to improve post-operative morbidity. The methods used in this mapping procedure have remained unchanged for 60 years. The procedure requires passing current into the patient's brain to temporarily disrupt function and is associated with problems such as seizure induction during surgery. Further, the procedure typically takes 45 minutes, is not well tolerated in many patients and must be aborted in up to 25% of cases. Hence, a new method to map language is needed. The goal of this proposal is to assess whether high gamma activity (70-250 Hz) in the human electrocorticogram (ECoG) can be used to accurately and rapidly track critical language cortex in the operating room. It is hypothesized that high gamma activity, a likely index of cortical activation, can aid surgeons identify critical language areas and potentially replace classic intra-operative ESM eliminating the risks associated with this procedure and shortening the surgical time. In order to investigate this issue, two studies will be conducted. First, we will record ECoG in the awake neurosurgical patient in the operating room during the picture naming/number counting study traditionally employed by neurosurgeons. Second, we will perform a brief verb generation study in the OR in these same patients. It is predicted that during picture naming and number counting regions showing arrest during ESM will elicit high gamma activity. We further predict that the high gamma map of language obtained in the verb generation task will also correlate with the stimulation mapping results obtained during study one. If obtained the results would provide evidence that language mapping using ECoG and high gamma recording might provide a safer and faster alternative method to the traditional intra-operative ESM method resulting in reduced intra-operative morbidity and improved post-surgical outcome. Neurosurgeons need to identify critical language areas in the human brain to avoid causing a language problem when removing tumors and other brain abnormalities. The current method to accomplish this requires a 60 year old approach that has serious side-effects such as causing a seizure in the operating room. We propose to apply a novel method of brain electrical analysis to replace the traditional method employed. If successful the results would lead to safer and more effective neurosurgical procedures. [unreadable] [unreadable] [unreadable]

It is widely accepted that human cognition is supported by distributed neural networks engaged in a task dependent manner. Prefrontal cortex (PFC) is crucial to top-down control of such networks across multiple cognitive domains including attention, language and memory. Despite the widespread use of the term "topdown" control there is surprisingly little neurocognitive research on how or where this might be instantiated at a neural level. For instance, it is unknown whether excitation vs inhibition of activity in cognitive tasks represents a unitary gain control mechanism, such that activity is either up or down regulated along one continuum by one control mechanism or whether top-down control reflects the net activity of multiple and distinct RFC-dependent excitatory and inhibitory mechanisms. We propose that PFC exerts independent excitatory and inhibitory control to support goal-directed behavior. To address this hypothesis we will perform a series of experiments examining whether top-down is supported by distinct excitatory and inhibitory mechanisms with independent time courses. We will further assess whether top-down control is supported by excitement or inhibition of all of a certain class of stimuli based on pre-set instructions (Task Set top-down control) or whether PFC top-down control can also rapidly exert excitation and inhibition in early perceptual regions on a trial by trial basis (Trial-by-Trial top-down control). To address these questions we will perform fMRI studies in normal controls, electrocorticography studies (ECoG) in patients with implanted subdural electrodes and lesion studies in patients with focal PFC damage centered in mid-lateral PFC (BA 44/9/46). The ECoG studies will focus on the newly described high gamma response (70-250 Hz) that we have shown to be a robust neural marker of cortical activation. The patient lesion studies will be coupled with parallel TMS inactivation studies of mid-lateral PFC in normals. The lesion and TMS studies will assess whether the proposed lateral PFC activations observed in the ECoG and TMS studies are critical to task performance. If successful the results of this project will provide crucial data using converging methods on the role of PFC in top-down control. The proposed work has broad theoretical implications and will also inform translational physiologically based interventions for neurological and neurosurgical patients with PFC damage.

Although we often think of emotion as the foe of reason, scientists have recently discovered that emotions are sometimes helpful for decision-making. For example, if the car in front of you suddenly brakes, you may feel a jolt of fear. The feeling of fear signals that something must be done quickly to avoid harm. On the other hand, unwarranted fear may motivate the decision to avoid an otherwise rewarding situation. For example, a shy student may not benefit from a class discussion because they are too afraid to speak up. How does the brain compute which emotions are helpful for decision-making and when the influence of emotion should be suppressed or integrated? The goals of this project are to understand how the brain represents (a) helpful and hurtful emotions and (b) the incorporation and inhibition of emotional influences on decision-making. With support from the National Science Foundation, Jennifer Beer at the University of Texas and Robert Knight at the University of California, Berkeley will address these questions by conducting parallel neuroimaging (functional magnetic resonance imaging) studies of healthy individuals and behavioral studies of patients with brain damage in regions that are hypothesized to be involved in emotion and inhibition (i.e., specific subregions of the frontal lobes). The studies will present human volunteers with emotional stimuli that are designated as helpful or hurtful for a subsequent decision. The studies systematically assess the influence of emotions on a number of computations that support decision-making: perceptions of risk, attention, and how deeply the decision is considered. The neuroimaging studies will test for brain regions that represent (a) helpful and hurtful emotion and (b) the incorporation and inhibition of emotional influences on decision-making. The studies of patients with brain damage will test whether the damaged region is necessary for these processes.

This work will result in a more comprehensive understanding of how the human brain represents emotional information and its role in decision-making. This research may lead to new models of decision-making which would have significant implications for disciplines beyond cognitive neuroscience such as economics and psychology. The funding from this research will be used to support research training opportunities for undergraduate and graduate trainees in social psychology, cognitive neuroscience, and brain imaging. The project results will be disseminated through publications and presentations to scientific and lay audiences.

DESCRIPTION (provided by applicant): Neuropsychological, neuroimaging and electrophysiological research supports a crucial role of lateral prefrontal cortex (PFC) in executive control of human behavior. Altered PFC function underlies a host of debilitating developmental, neurological and psychiatric disorders. Despite significant progress in PFC research, the real-time processes supporting PFC control of human cognition remain undetermined. We propose that PFC uses oscillatory dynamics to implement cognitive control of task-dependent neural networks. To address this hypothesis we will use a unique combination of direct cortical recording in neurosurgical patients (Electrocorticography; ECoG) with superb spatio-temporal resolution and signal-to-noise ratio, along with cutting-edge analysis to define PFC-dependent neural networks supporting cognition. We will examine three tasks probing hallmarks of frontal lobe function (categorical perception, working memory and controlled vs. automatic attention) across three sensory modalities (audition, vision, touch) to address the generalizabilty of PFC dependent oscillatory network control. We will then examine the causality of the oscillatory dynamics observed with ECoG using scalp EEG and behavioral methods in patients with focal lesions centered in lateral PFC. Aim 1 will use ECoG recording to define the role of PFC control of phase locking of low frequency neural oscillations between task dependent networks and whether cross frequency coupling (CFC) between high gamma oscillations (60-200 Hz) and low frequency oscillations tracks task performance. Aim 2 is the most exploratory and draws from a recent finding in our laboratory demonstrating that an exquisite cortical microstructure observed at 4 mm cortical resolution supports categorical perception of phonemes in the temporal lobe. Using ECoG we propose to examine whether such a sub-centimeter microstructure extends to the PFC during cognitive control. Aim 3 wil examine the causality of each oscillatory process in patients with either focal lateral PFC or posterior parietal cortex (PPC) damage. We predict that deficits in neural oscillations in the lesioned hemisphere wil correlate with behavioral deficits and reveal novel patterns of neuroplasticity in the non-lesioned hemisphere. In summary, the proposed work will define the role of PFC in both distributed and local network function in the support of human behavior.

DESCRIPTION (provided by applicant): Neuropsychological, neuroimaging and electrophysiological research supports a crucial role of lateral prefrontal cortex (PFC) in executive control of human behavior. Altered PFC function underlies a host of debilitating developmental, neurological and psychiatric disorders. Despite significant progress in PFC research, the real-time processes supporting PFC control of human cognition remain undetermined. We propose that PFC uses oscillatory dynamics to implement cognitive control of task-dependent neural networks. To address this hypothesis we employ a unique combination of direct cortical recording in neurosurgical patients (Electrocorticography; ECoG) with superb spatio-temporal resolution and signal-to-noise ratio, along with cutting-edge analysis to define PFC-dependent neural networks supporting cognition. We will examine three tasks probing hallmarks of frontal lobe function (categorical perception, working memory and controlled vs. automatic attention) across three sensory modalities (audition, vision, touch) to address the generalizabilty of PFC dependent oscillatory network control. We will then examine the causality of the oscillatory dynamics observed with ECoG using scalp EEG and behavioral methods in patients with focal lesions centered in lateral PFC regions activated in ECoG recording. Aim 1 will use ECoG recording to define the role of PFC control of phase locking of low frequency neural oscillations between local modules of task dependent networks and examines whether phase locking and cross frequency coupling (CFC) between high gamma oscillations (60-200 Hz) and low frequency oscillations tracks task performance at the single- trial level. Aim 2 examines the causality of ECoG defined oscillatory processes in patients with focal lateral PFC or posterior parietal cortex (PPC) damage. Finally, we predict that deficits in neural oscillations in the lesioned hemisphere will predict behavioral deficits and that neural oscillations in the non lesioned hemisphere will reveal novel patterns of neuroplasiticity and behavioral compensation. In summary, the proposed work aims to define the role of PFC in both distributed and local network function in the support of human behavior with implications for understanding normal as well as disordered brain function.

DESCRIPTION (provided by applicant): The visual environment contains more information than can be processed simultaneously. Due to this limited processing capacity of the visual system, it is necessary to select the behaviorally most relevant information for further processing and to filter out the unwanted information, a fundamental ability known as attentional selection. There is converging evidence from physiology studies in monkeys and neuroimaging studies in humans that attentional selection occurs at multiple stages along the visual pathway and is controlled by a network of higher-order areas in frontal and parietal cortex that includes the frontal eye fields (FEF) and the lateral intraparietal area (LIP) in the monkey and functionall similar areas in the human. In monkeys, physiology studies have begun to characterize the interactions across the network by simultaneously recording from two or more interconnected nodes of the attention network. One important result of these studies suggests that the strength of attentional modulation depends on the degree of neural synchrony between areas. In contrast, in humans, little is known about the temporal dynamics and functional interactions across areas of the attention network. Further, despite the macaque brain serving as prime model for a basic understanding of human brain function, it remains unclear how neural mechanisms related to perception and cognition compare across primate species. By recording intracranially from frontal and parietal cortex of monkeys and of epilepsy patients, who are chronically implanted with subdural grids for diagnostic purposes, while performing an identical spatial attention task we pursue two main goals in this project: (i) to characterize the temporal dynamics of the human attention network; and (ii) to compare electrophysiological signals related to spatial attention and obtained in functionally similar areas across primate species (monkey/human). The central hypothesis is that modulation of oscillatory activity plays an important functional role in spatial attention control and can predict behavioral outcome in both primate species. The significance of the proposed research is that it will contribute to our understanding of a fundamental cognitive operation, selective attention, the impairment of which has devastating consequences on human health. Attentional deficits are frequently observed in neurological diseases, e.g. after stroke, leading to visuo-spatial neglect, an impairment in directing attention to contralesional visual space, as well as in psychiatric diseases (e.g. schizophrenia). In addition, our proposed studies will be the first to directly compare human and macaque physiology, thereby connecting two different bodies of literature, i.e. EEG/fMRI and macaque electrophysiology.

? DESCRIPTION (provided by applicant): The natural environment is cluttered with stimuli and the brain has limited processing capacity. Attentional mechanisms are therefore needed to guide the selection of behaviorally relevant information. The present application is a competing renewal submission for our project Neural basis of visual attention (R01- MH64043-10). Work during the previous funding period used fMRI to characterize attention-related functions at multiple stages of the human attention network, including the thalamus and the fronto-parietal network. The present application extends this work, using electrophysiological recordings in humans and monkeys to investigate temporal dynamics and functional interactions across the attention network (both cortex and thalamus). Specifically, we aim to characterize the neural basis of object-based selective attention. Objects are typically the units of selection. When attention is spatially allocated to part of an object, an `attentional spotlight' (and the facilitaed processing associated with it) automatically expands to match the extent of the object's boundaries. That is, attention spreads to include task-irrelevant locations within the object. Classic attention theories assume that a unitary and indivisible spatial mechanism mediates such object-based selection. However, recent evidence challenges this characterization. First, the attentional spotlight, rather than being sustained, flashes rhythmically, sampling the visual environment at frequencies in the theta band (4-8Hz), with alternating temporal windows of relatively enhanced and diminished processing. Second, our behavioral studies from the previous grant cycle support the existence of two spatial mechanisms, concurrently sampling the visual scene: (i) a fixed spotlight that rhythmically samples the most relevant location, and (ii) a moving spotlight that rhythmically monitors less relevant locations. The present project wil challenge classic models of attention by investigating the neural basis of rhythmic selective processing within the framework of object- based attention. By recording from multiple nodes of the attention network in humans and macaques (using the same attention task for both species), we will probe the central hypothesis that attentional selection is a highly dynamic process that operates concurrently at multiple locations. In humans, we will obtain intracranial recordings in epilepsy patients (i.e., electrocorticography [ECoG]). In macaques, we will simultaneously record from interconnected regions of FEF, LIP, area V4, and the pulvinar. Specifically, we will (i) investigate a dissociation of function between FEF and posterior PPC, (ii test the functional role of the pulvinar in coordinating temporal dynamics across cortical network nodes and (iii) relate neural signals, including oscillatory activity, to rhythmic behavior. Impairments of attentional selection have devastating consequences on human health (e.g., after stroke, and in diseases, such as schizophrenia). The significance of the proposed research is that it will probe potentially paradigm-shifting hypotheses using an innovative approach that combines cutting edge neuroimaging techniques with intracranial electrophysiology in two primate species.

Project summary/abstract The research proposal focuses on studying the electrophysiological and oscillatory mechanisms underlying decision-making involving risk-reward tradeoffs. Specifically, we will record electrophysiological data from patients with extensive prefrontal cortex ECoG coverage (tens to hundreds of electrodes in lateral PFC, orbitofrontal cortex, and other PFC areas) while they carry out a gambling task. Using this data, we will test the hypotheses that oscillatory mechanisms reflect local valuation and global top-down control processes in decision-making. Decision-making is disturbed in numerous psychiatric disorders including schizophrenia, major depression, and a variety of personality disorders. As such, a deeper understanding of the cortical mechanisms supporting decision-making capacity has the promise to shed new light in a host of disorders relevant to the mission of the NIMH.

How do we extract salient information from an ever-changing and noisy environment? Project 3 addresses this fundamental question in perception using direct brain recordings in humans (electrocorticography; ECoG) to assess two models of sensory acquisition. The Active Sensing model posits that high-level inputs act to rhythmically sample the sensory word and filter out noise. The related predictive coding model theory posits that prior knowledge enhances perception with the brain making predictions about upcoming stimuli to sharpen low-level sensory processing. We propose that both processes share similar neural substrates - neuronal rhythm-based engagement of frontal, premotor, motor and sensory cortical networks to enable active and predictive sampling of the world to enhance perception. We employ ECoG to measure neural oscillations and high frequency activity (HG; 70-200 Hz; surrogate for intracortical SUA activity) and employ network analysis approaches to define the role of top-down control of active sensing and predictive coding in the human brain. Two or our proposed human ECoG studies are performed in monkeys in Project 4 permitting a rich inter-species comparison of the neural substrates of sensory acquisition. AIM 1 tests the hypotheses that motor/premotor systems control auditory sampling rhythms and actively suppress distracting information. This aim also explores whether lateral prefrontal regions provide additional control to the motor/premotor-auditory active-sensing network. AIM 2 addresses how prior knowledge enhances speech perception and `fills-in' degraded speech representations in auditory cortices. Given the use of speech stimuli this study will only be performed in humans. This Aim directly tests the predictive coding model and examines if similar neural substrates support both predictive coding and active sensing. AIM 3 compares our ECoG data to the laminar LFP/CSD and MUA profiles and network parameters obtained in parallel monkey auditory Project 4. These unique cross-species data will be used to identify the cell populations and physiological processes that generate ECoG components in monkeys and humans providing unprecedented insights into cortical physiology in humans. We predict that active sensing mechanisms are modality independent and will also compare our finding from the auditory monkey-man to the visual human and monkey active sensing studies in Projects 1 and 2. Core C provides critical DTI and resting state fMRI to correlate with our ECoG network and HG data and Core B provides for data standardization and sharing. Finally, Project 5 provides the computational and modeling infrastructure necessary to build and refine cell and systems level models of the world is sampled. Active sensing and predictive coding are likely impaired in a host of disabling psychiatric, neurological and developmental disorders making the understanding of these processes central to the mission of the NIMH.

PROJECT SUMMARY: The ability to learning rapidly is one of the defining features of human cognition. Despite its importance, the circuit mechanism that governs rapid learning in humans is unknown. It has been proposed that prior-knowledge or a ?mental schema? facilitates rapid learning via prefrontal-hippocampal network interactions to improve acquisition of novel associative memory. There is, however, limited empirical evidence supporting this model of learning. Moreover, comparisons of circuit dynamics underlying rapid learning have not been conducted between humans and nonhuman primates. The proposal bridges systems neuroscience across primate species and addresses three fundamental knowledge gaps: 1) Circuit dynamics between the prefrontal cortex and hippocampus that support associative and categorical learning, 2) The influence sleep overnight on memory retention, 3) Commonalities and differences in neural activity and circuit dynamics between human and nonhuman primates during learning. To establish cross-species comparisons, we will conducts a set of experiments in humans tightly linked to the nonhuman primate projects to elucidate the circuit mechanisms of cortical-hippocampal interactions during rapid schema-based and categorical learning. The pre-surgical evaluation of patients with epilepsy provides a unique and potent opportunity to study these brain networks directly. Specially, we will use large-scale high-density intracranial electrodes to record neural signals from prefrontal cortex and hippocampus while patients perform associative and categorical learning. We will also leverage the unique ability to record single neurons in the human hippocampus and medial prefrontal regions to directly compare neural activity across species. Our studies will greatly advance the neurobiology of learning and memory, for which impairments form core clinical features of diverse neurological disorders such as Alzheimer's disease, autism, major depression, and epilepsy. Understanding the neural mechanisms of rapid learning will provide critical framework to develop circuit specific intervention in people with disordered memory.