Bijan Pesaran - US grants

DESCRIPTION (provided by applicant): The long-term goal of this research is to understand the neural representations that guide speech production and perception. While speech and language are central features of human behavior, the neural mechanisms of speech processes remain poorly understood. We will build on recent advances in the neural control of visually-guided reaching to develop analogous models for speech processing. Aim 1 will employ a pseudo- word repetition identification paradigm to test the hypothesis that speech production is encoded in neural activity using an auditory-articulatory representation. Aim 2 will employ a pseudo-word identification paradigm to test the hypothesis that speech perception is encoded using an auditory-articulatory representation. Aim 3 will attempt to replicate findings in Aims 1 and 2 employing a complementary set of pseudo-words. Neural activity will be directly recorded using intracranial electrodes in human patients with pharmacologically- intractable epilepsy. I predict that neural activity in specific cortical areas supporting speech and language will exhibit invariances for particular auditory-articulatory mappings. This will be evidence that speech is processed jointly in auditory space and articulatory space and could explain the specific role of auditory signals in speech production and articulatory signals in speech perception. PUBLIC HEALTH RELEVANCE: Augmentative and alternative communication systems give people with severe communication disorders the ability to communicate. Even after a severe loss of motor function, people can maintain the ability to express themselves with language if they have a communication channel. The long-term goal of this research is to give people with severe motor disorders the ability to speak and communicate by decoding the neural representations that support speech and language processing.

We move our eyes not only to see the world better but also to guide our movements in the world more accurately. Often we are unaware of how we use our eyes to help us see and move better, because the way we coordinate our eye and hand movements is so dynamic and flexible. During coordination, many areas of the brain become active, and models predict that different areas in the brain must communicate with each other in order to coordinate our eye and hand movements. With funding from the National Science Foundation, Dr. Bijan Pesaran is pursuing two major research goals that develop eye-hand coordination as a model for understanding how brain areas work together during cognition. Flexible communication between different brain areas is fundamentally important to a wide range of cognitive processes but very little is known about how brain areas talk to each other to guide our behavior. The experiments carried out under this project investigate how the coupling between coordinated eye-hand behaviors depends on cooperation and inhibition between visual saccade and reach representations. In one set of experiments, Dr. Pesaran is examining how cooperation between reaches and saccades leads to a common stage of target selection before coordinated movements. In a second set of experiments, Dr. Pesaran is examining how during reaching, the selection of new eye movements is inhibited by reaching. Inhibition slows down new eye movements, so that they occur after the reach movement is finished.

Eye-hand coordination involves high-level cognitive processes, and understanding coordination can serve as an important model for understanding how brain systems interact to carry out cognitive processes in general. The interdisciplinary research program also inspires an integrated educational plan to increase awareness of neuroscience amongst students and increase the use of quantitative tools in neuroscience research. An increase in the level of quantitative and statistical sophistication is a hallmark of the next generation of brain scientists. This increase is needed to successfully tackle challenging research questions about how brain areas cooperate to guide behavior. Outreach to increase awareness of neuroscience is extending beyond the university curriculum into K-12 programs. Dr. Pesaran is demonstrating neuroscience concepts in a school outreach program in public schools. The educational outreach program aims to improve awareness of neuroscience amongst school children and encourage them to explore neuroscience educational opportunities in the future.

DESCRIPTION (provided by applicant): Vision is critical for effective control of action. The dorsal stream projects to the posterior parietal cortex (PPC), and transforms visual input into corresponding responses to guide movements of different effectors. An influential model proposes that the dorsal stream is organized into dedicated visuomotor channels that are effector-specific and operate independently. Experiments for this proposal seek to rigorously test this model against alternatives in which channels guide movements of multiple effectors and can interact. We will first test whether coordinating visually-guided behaviors recruits effector-specific processes by recording neurons in the parietal reach system (Aim 1) and the parietal saccade system (Aim 2) during coordinated behavior. We will then test whether channels operate independently or interact by measuring the strength and functional significance of cross-areal coherence between the parietal reach and saccade systems (Aim 3). Despite the central role of coordination to visual behavior, how visual space for reaching and eye movements influence each other and whether cross-areal interactions within PPC underlie this influence is not known. Understanding the mechanisms of coordinated visual behavior will help us understand how visual behavior arises from interactions between different visual representations more generally. Damage to the PPC leads to optic ataxia, apraxia and neglect, is debilitating, and is suffered by many patients each year in the US with significant social costs Understanding how the brain constructs visual-spatial representations underlies the effective treatment of patients with posterior parietal damage.

Abstract The Design and Fabrication Module supports vision research in the at New York University by equipping and sustaining a high-quality central resource capable of the expert design, timely construction, and efficient repair of specialized electronic, mechanical, and electromechanical devices for research. The Design and Fabrication Module provides moderate or extensive support for 13 members of the Vision Core, including 3 young investigators and 10 NEI funded investigators, 7 of whom hold qualifying grants.

To distinguish parts of our body from other objects around us, the brain needs to build an internal image of our body by merging information from the skin, muscles and joints with information from our eyes. This project is aimed at characterizing how we build our sense of self, by using multisite brain recordings and virtual reality technologies. The results will impact national needs in the consumer, healthcare, military, and industrial settings by advancing the fundamental engineering and neuroscience knowledge necessary to create the next generation of brain-machine interfaces, which are envisioned to support the integration of artificial and natural sensory information. Optimizing these systems depends critically on understanding how natural sensory signals interact within and among brain areas, a knowledge gap that directly addressed by this project. The educational goals associated with the project are designed to advance discovery and understanding of engineering and neuroscience, while also promoting teaching, training, and learning beyond the regular bounds of these disciplines. These goals are achieved by: 1) Engaging high school students underrepresented in STEM fields through the development of hands-on instructional modules, 2) Promoting interdisciplinary undergraduate research opportunities via internships at Arizona State University, 3) Mentoring students in the broader implications of scientific research through exposure to organizations engaged in the ethical, societal, and policy implications of neuroscience research, and 4) Engaging the public in scientific discourse through public lectures and exhibits, and thereby promoting broad dissemination of the work to enhance scientific and technological understanding.

Estimating the state of the body through the integration of available sensory cues (multimodal state estimation) is a critical integrative function for most organisms. Although much is known about state estimation for the upper limb at the behavioral level, the underlying neural mechanisms remain poorly understood in cortical areas, particularly at the network level. This is due to several factors: 1) the cortical areas believed to play a role in limb state estimation are heterogenous with regard to the relative strength of their sensory inputs and display both multisensory enhancement and suppression depending on context; 2) technical limitations mean functional interactions among these areas have been challenging to characterize; 3) the relation between sensitivity to visual and somatic cues and prevailing computational theories of multisensory integration have been incompletely explored; 4) multimodal areas are thought to contribute to both perceptual and action-based body representations but how these representations interact at the neural and behavioral levels is not well understood. As a result, it is unclear how a coherent multimodal estimate of the state of the upper limb is constructed and maintained. The proposed series of studies address these issues by quantifying changes in neural spiking, local field potentials, and neural coherence within and among fronto-parietal areas of the monkey implicated in state estimation using virtual reaching tasks that alter the reliability and semantic information of visual cues.

Abstract In this proposal, we will develop next-generation flexible micro-electrocortigraphic (µECoG) and penetrating electrode arrays using active electronics in complementary metal-oxide-semiconductor (CMOS) technology. Active electronics enable amplification and multiplexing directly at each electrode, eliminating the need for implanted electrodes to be individually wired to remote electronics and greatly increasing the number and density of electrodes that can be recorded and stimulated. The flexibility of our arrays allows them to conform to the irregular geometry of the brain, yielding higher fidelity signals and reduces damage to the brain when used in penetrating configurations. Integrated wireless data and power enables completely tether-free implants. Together, these innovations enable us to take high resolution measurements over large areas of the brain while being less invasive, a substantial improvement over the current state-of-the-art. In surface recording structures, we will demonstrate electrode arrays of up to 65,536 electrodes and amplifiers, spaced just 25.4µm apart, where each electrode can be simultaneously sampled at 20 ksps, enabling a cellular-resolution brain interface across a 64 mm² brain area. Each electrode can also be independently stimulated, or stimulated with patterns of activation, mimicking more natural excitation patterns. In penetrating arrays, we will demonstrate fully integrated, flexible penetrating neural probes with up to 512 electrodes per shank. The probe ?head? containing active electronics will fold over the outer surface of the cortex, at the point of the probe?s insertion, positioning its inductor for a near-field link through the skull. This link will be powered wirelessly with near-field radio-frequency data telemetry, eliminating the need to run wired interconnections through the skull. Integration with wireless interfaces will permit sealing chronically- implantable probes subcutaneously and in a manner in which the entire probe floats on the brain. The developed technologies will be rigorously tested in vitro and in vivo. This project will make high density electrode arrays based on manufacturable flexible CMOS technology available for the broader neuroscience community, enabling studies of large-scale recording and modulation in the nervous system. The innovations generated through this work have the potential to revolutionize our ability to understand the brain, and will improve epilepsy surgery outcomes as well as advance the performance of motor and auditory prosthetics. This project leverages a successful, long-term collaboration between clinicians, engineers, material scientists and neuroscientists at Duke University, Columbia University, New York University and the University of Illinois at Urbana-Champaign, to translate active, flexible electronics technology into next generation implantable neurological devices.

This proposal presents an academic-industrial partnership to develop an instrument for 2-photon laser scanning microscopy (2p-LSM) in non-human primates (NHP) for transformative basic science and clinical translation. Our team combines academic strengths in neuroscience and neural engineering at New York University (PI Pesaran, co-I's Movshon and Peron) and industrial engineering at ThorLabs (led by Jeff Brooker, Vice-President of Life Sciences) to translate recent advances in 2p-LSM imaging by Karel Svoboda and his group at Janelia Research Campus to the monkey. Many researchers who study the primate brain would like to add 2p-LSM to their experimental toolkit. Unfortunately, no existing microscope can be used for NHP work without significant customization due to the size and shape of the primate brain. We plan to design an instrument that will overcome these challenges and serve a community of researchers. Recently, a 2-photon random-access mesoscope, or 2p-RAM, has been developed. The 2p-RAM has the same spatial resolution as a standard 2p-LSM but uses a novel optical strategy to scan a 5 mm field of view, up 1 mm deep. The current 2p-RAM is designed for use in the mouse. To adapt it for NHP use, we propose to work with Thor Labs to develop a manipulator that can position the objective lens with six degrees of freedom over any optically accessible part of the cortical surface. We expect the new instrument to provide: 1- Random access scanning to image populations of neurons selected for study. In the mouse, roughly 10,000 neurons can be selected from larger populations and their activity imaged simultaneously; 2 - Motorized control to position and reposition the objective at micron spatial resolution and micro-radian angular resolution. This will allow users with different target areas to share an instrument; 3 - Dynamic access scanning to move the objective quickly and precisely to image multiple 5-mm fields of view. This will make possible integrated measurements of populations of neurons in many different brain regions; 4 - Software to control the instrument and acquire the data. In Aim 1, we develop a motorized platform for NHP-optimized imaging using the microscope and optimize the associated surgical procedures. In Aim 2, we develop a motorized platform for NHP-optimized imaging using the 2p-RAM. In each Aim, we develop software to control the instrument and acquire the data it will generate. At the end of the project, Thor Labs will develop, build, and sell the technology commercially. In addition, we will disseminate the technology openly, including documentation for instrument fabrication and user training. A key strength of our proposal will be a community of users with stated interest in using the device and we provide many letters of support. This instrument will be a major advance for research on the primate brain and will drive advances in human health.

Project Summary During behavior, the oculomotor system is tasked with selecting objects from an ever-changing visual field and guiding eye movements to these locations. The attentional priority given to sensory targets during selection can be strongly influenced by external stimulus properties (?bottom-up?) or internal goals based on previous experience (?top-down?). Although these exogenous and endogenous drivers of selection are known to operate across partially overlapping time scales, how neural circuits mechanistically support top-down and bottom-up processing has been difficult to disentangle. This is because the neural circuits for spatial attention and selection are distributed across the frontal and parietal cortices and operate across multiple spatial scales spanning the activity of individual neurons and neuronal populations. In this Targeted Brain Circuit R01 Project proposal, an experimental group (Pesaran/NYU) and a theory group (Shanechi/USC) will use cutting-edge techniques developed under the NIH BRAIN Initiative support to validate predictive models of neuronal dynamics and test hypotheses about how frontal-parietal cortices perform attentional selection. A behavioral task that dissociates bottom up and top-down processing will let us define bottom-up and top-down target states. We will then build predictive models of neuronal dynamics within and between frontal and parietal cortex and empirically validate the models by stimulating neural activity to achieve the desired neural state. Aim 1 validates predictive models of local circuit dynamics. We will stimulate within PFC to achieve target states in PFC. Aim 2 validates predictive models of long-range circuit dynamics. We will stimulate sites in PPC that functionally connect to PFC in order to achieve target states in PFC. Aim 3 validates predictive models of distributed circuit dynamics. We will simultaneously stimulate both PFC and PPC to achieve the target states. In each case, successfully directing activity toward the target state will indicate the model is valid. If the target state reflects a causal role in attention, as opposed to correlating with attentional processes, we predict that behavioral choices will be biased. This proposal tackles several of the major topic areas of the BRAIN 2025 report. We will identify fundamental principles about circuit dynamics and functional connectivity for understanding the biological basis of mental processes through development of new theoretical and data analysis tools (Topic 5). We will produce a dynamic picture of the functioning brain by developing and applying improved methods for large-scale monitoring of neural activity (Topic 3). We will demonstrate causality by linking brain activity to behavior with precise interventional tools that change neural circuit dynamics (Topic 4). Recent years have seen dramatic advances in our ability to experimentally interface with the primate brain with increasing precision scale. A fruitful interplay between multiscale experiments and predictive modeling that we propose will let us test hypotheses about how flexible behaviors are controlled by large-scale neural circuits.

Project Summary This proposal explores an emergent computational framework for understanding the neural population codes that support flexible, context-dependent behavior. The current state of the field is based on two competing views. According to the circuits view, fixed behaviors arise from specific anatomically or genetically defined cell populations that serve specific functions. Alternatively, the network computation view instead holds that neural activity provides mixed representations of task variables and can be understood only based on the joint activation of many neurons. Currently, these competing views are pursued by different communities with different tools, different behavioral paradigms and different model organisms. This has led to a disconnect between neural computation and the underlying biological circuit mechanisms. Here we propose a unified framework, in which the combinatorial activity of biologically-identified populations of neurons shapes the computations through low dimensional dynamics. A new interdisciplinary team of investigators - Pesaran (NYU - primate experimentalist), Johansen (RIKEN - rodent experimentalist) and Ostojic (Ecole Normale Superieure - computational theory) will develop the computational theory and apply it to flexible input-output tasks in multiple species - rats and non-human primates. To achieve these goals, we will analyze recurrent neural network models trained to perform a sensory-motor context-dependent decision-making task and fit low-dimensional models to experimental data. We will perform experiments in rodents and non-human primates to validate model predictions that combinatorial coding can support context-dependent behavior and is behaviorally-significant (Aim 1). We will analyze whether biologically-defined cell types map onto computationally-defined cell classes during context- dependent behavior by determining how activity in combinations of genetically and anatomically identified PFC cell types corresponds to low-dimensional dynamics (Aim 2). In parallel, we will determine how biologically- defined cell types control context-dependent behavior during explicitly-cued and implicitly-signaled contexts by performing optogenetic perturbations of anatomically-defined PFC cell types and testing behavioral performance (Aim 3). Successful completion of these aims will link neural computation to biology across multiple species to deliver a computational framework explaining how biological circuit mechanisms give rise to neuronal computations that mediate context-dependent behavior.

Everyday tasks, such as tying shoelaces, throwing a ball, reaching for a cup involve performing complex, require precise and coordinated hand and arm movements in various behavioral contexts. How our brain controls these rich movements; that is, the neural basis of motor control, remains elusive. Investigating this neural basis not only will advance our understanding of brain function, but also will have broad translational implications for developing brain-machine interfaces to restore function in disabled patients and for understanding neurological disorders such as Parkinson’s disease and devising deep brain stimulation therapies. This proposal investigates how a population of neurons in the motor cortex of the brain controls movements by utilizing motor experiments that record simultaneously from large motor cortical areas including thousands of neurons, and by leveraging computational tools that can uncover behaviorally relevant structure in the complex neural population data.

This research involves a very close collaboration between computational and experimental methods. The experimental component incorporates a variety of different behavioral conditions that produce a rich repertoire of movements, while recording from large populations of motor cortical neurons in an animal model. The primary focus of this project is on the relationship between movement kinematics (the trajectory of motion) and the temporal dynamics of these large neuronal populations in motor and premotor cortex. To study this relationship, we analyze the computational component that characterizes neural population dynamics using a low-dimensional state variable that is directly related to behaviorally relevant movement kinematics. This is achieved by leveraging linear dynamical modeling methods that can identify low-dimensional behaviorally relevant state dynamics in motor cortical populations, examining how these states change, and modeling local and input dynamics during behavior. Integrating these computational and experimental components, will advance our understanding of the dynamical principles underlying how motor cortical population activity gives rise to rich movements and has the potential to lead to breakthroughs in restoring motor function in patients with impaired movement.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.