Reza Shadmehr - US grants

We aim to study the changes that take place in the output of the motor system and its neural correlates in the brain as humans learn a motor skill with the arm, and describe how the neural and functional representation of the skill changes in the hours after completion of practice. We hypothesize that during these hours, motor memories undergo a process of consolidation. Our principal experimental tool is a robotic system that subjects learn to control with their arm. We will perform psychophysical and neuroimaging experiments coupled with computational modeling to pursue the following objectives: To understand, using a mathematical model, the changes that take place in the motor output as reaching movements are made in a novel force field; To describe the time course of changes in functional properties of motor memory in the hours after completion of practice; To describe the neural correlates of motor memory consolidation through functional imaging; To ask whether the time- dependent functional properties associated with motor memory are intact in amnesia; To better understand the role of the cerebellum and the striatum in learning of reaching movements by studying motor learning in patients with cerebellar disease and Huntington's Disease. With a better understanding of how the normal brain learns a new motor skills and how time influences its representation, we will better understand motor learning and rehabilitation processes in the damaged brain.

[unreadable] DESCRIPTION: Our long term goal is to construct a unifying framework that explains the roles of the basal ganglia and the cerebellum in control of saccades and reaching. We suggest that a fundamental assumption regarding control of movements, the idea that variability is mostly because of noise, is incorrect. We show that there is systematic variability in the motor commands that initiate a movement to a target, and propose that this variability is a reflection of a systematic reduction in the internal value that the brain associates with a repeating stimulus. We hypothesize that this internal value is computed in the striatum. If this variability was uncompensated, that is, if movements like saccades were "open loop", then the variability would affect saccade endpoints. In healthy people, however, saccade endpoints are immune to this variability. We suggest that this is because control of movements is strongly dependent on internal models through the cerebellum, monitoring the outgoing motor commands and effectively "steering" to compensate for variability in the outgoing motor command that would lead to unacceptable inaccuracy, i.e., dysmetria. The compensation is effective only if this internal model is calibrated, which links the problem of control with adaptation. We propose a single principle of control and adaptation for both the saccadic and reaching systems: each is supported by a fast adaptive system with poor retention, and a slow adaptive system with strong retention. Expected costs and rewards of a movement are evaluated by the basal ganglia, resulting in an internal value that affects the motor commands that initiate the movement. As the motor commands are generated, the cerebellum monitors them and predicts their sensory consequences, producing adjustments that "steer" the movement to the goal. We propose that adaptation is faster in the mechanism that steers the movement (cerebellum for both reaching and saccades) than the mechanism that initiates the movement (motor cortex for reaching, superior colliculus for saccades). PUBLIC HEALTH RELEVANCE Our hypothesis is a new, coherent theory of how various brain structures like the basal ganglia and the cerebellum contribute to control of voluntary movements like saccades and reaching. While cerebellar patients have been consistently impaired in motor learning, our hypothesis presents a potential solution to how rehabilitation may proceed in these patients to help their recovery. The role of basal ganglia in control of movements has remained a deep puzzle. Our hypothesized link between this structure and the internal value of action may help early diagnosis of diseases of the basal ganglia through experiments that quantify changes in trajectories of the eyes and the arm in response to changes in value of the stimulus that affords these movements. [unreadable] [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Adaptation is central to biological motor control because our muscles, our motor plant, and the environment that we interact with have time varying properties. Some of these properties change suddenly and temporarily while others change slowly and may last a long time. We suggest that the way that the brain learns motor control is to a great extent a reflection of these time scales of change. In effect, we propose that the brain treats adaptation as a statistical problem that includes prior information about timescales of change. Our new approach solves two major problems which have significantly curtailed our current understanding of the neural basis of adaptation: (1) Current models of movement adaptation focus almost exclusively on error. On every trial, the error is associated with the present context and the system adapts its internal model by changing parameter values or weights. However, there is ample evidence that behavior can change not only as a function of the magnitude of errors, but also as a function of when these errors are made. Our approach predicts the strong dependence on the temporal and contextual history of the training trials in adaptation experiments. (2) Traditional motor adaptation theories assume that the CNS chooses a desired trajectory, estimates internal model parameters through trial and error, and then produces motor commands that move the limb along this desired trajectory. However, movements have costs and gains that are often described in terms of their end result, not a specific trajectory. Indeed, close inspection of behavior during adaptation reveals clear deviations from the predictions of a desired trajectory assumption. Here we treat the problems of uncertainty about the world and our motor plant as well as cost functions in a single framework. We suggest a new theory to guide experiments in sensorimotor learning. What we are proposing here is a fundamental shift away from the current focus on motor error and desired trajectories - ideas that have been the mainstays of sensorimotor research. First, we link learning of internal models to a causal structure of how the body might be affected by real world perturbations. Next, we link changes in internal models to changes in sensorimotor control. As a result, we suggest that the adaptive behavior that we and others have measured is really a result of two concurrent computational processes: statistical formulation of internal models (a process akin to system identification), and the use of those internal models in the framework of optimal control to produce motor commands. Development of this new theoretical framework will ultimately shed light on the computations that are performed by structures in the brain that participate in control of our movements.

DESCRIPTION (provided by applicant): The Biomedical Engineering Doctoral training program at Johns Hopkins University aims to train talented Students from engineering and other quantitative sciences for careers in biological and medical research. Our program is based on more than 45 years of educational experience in Biomedical Engineering, and a collaborative research environment made possible by our strong presence in both the engineering and the medical schools. The program is interdisciplinary and interdepartmental in nature. Program faculty are drawn from a wide range of departments. This includes but is not limited to the department of Biomedical Engineering, Neuroscience, Otolaryngology-Head and Neck Surgery, and Radiology in the School of Medicine, and the departments of Applied Mathematics and Statistics, Chemical and Biomolecular Engineering, Electrical and Computer Engineering, Material Science, and Mechanical Engineering in the Whiting School of Engineering. The faculty are engineers, applied mathematicians, neuroscientists, physiologists, physicians, cell biologists, and molecular biologists with both experimental and theoretical/computational research programs. Our sponsored research base remains exceptionally strong with funding from diverse sources. Students are drawn mainly from the top engineering programs in the United States. This highly competitive national pool has allowed us to maintain very high standards of selectivity. The signature of our educational program is our commitment to provide outstanding training in both biology and engineering. Our students learn biology and physiology alongside medical students in their first year, and engineering and advanced mathematics in their second year and beyond. They have the freedom to do research in any laboratory in the University. This philosophy has yielded exceptionally productive students: the students who have graduated in the past 10 years have produced 660 peer- reviewed papers as a result of their PhD research. This is an average of 5.7 papers per student. Support is requested for 13 pre-doctoral trainees. The average duration of the program is 5.8 years. The core of the program is research training in the research laboratories of the Program faculty. RELEVANCE: A program is being developed at Johns Hopkins University to support and train graduate students with educational and research interests in the field of Biomedical Engineering. The program trains talented students from engineering and quantitative sciences for careers in biological and medical research.

DESCRIPTION (provided by applicant): It is commonly assumed that saccades are ballistic and stereotypical. Yet, there is structured variability in the motor commands that initiate a saccade to a target. We propose that this variability is partly a reflection of a systematic change in the internal value that the brain associates with a visual stimulus. That is, our brain assigns a value to targets of our eye movements, and this relative value modulates the motor commands that move our eyes. If this variability was uncompensated, that is, if saccades were open-loop, then the variability in the motor commands that initiated the saccade would affect saccade endpoints. In healthy people, however, saccade endpoints are accurate despite the fact that the motor commands that initiate the saccade are variable. We suggest that this is because control of saccades is strongly dependent on internal models through the cerebellum, monitoring the outgoing motor commands and effectively `steering'the saccade by adding motor commands late in the saccade's trajectory. The compensation is effective only if this internal model is calibrated, which links the problem of control with adaptation. Here, we propose a single principle of control for saccades: Expected costs and rewards of a movement are evaluated by the basal ganglia, resulting in an internal value that modulates the motor commands that initiate the saccade. As the motor commands are generated, the cerebellum monitors them and predicts their sensory consequences, producing adjustments that steer the movement to the goal. If successful, our project may produce a major shift in oculomotor research by linking variability in trajectory of saccades with expected rewards and internal models. These concepts have proven fundamental in understanding control of other movements like reaching. As a result, we may be able to produce a single conceptual framework for how the brain controls movements in general. PUBLIC HEALTH RELEVANCE: Our aim is to produce a new, coherent theory of how various brain structures like the basal ganglia and the cerebellum contribute to control of saccades. While cerebellar patients exhibit dysmetria in their saccades, our proposal suggests that dysmetria is not random, but related to the intrinsic value that the brain assigns to the visual stimulus. The role of basal ganglia in control of movements has remained a deep puzzle;our proposal suggests that there may be a link between this structure and the internal value that the brain assigns to a visual stimulus, and that this factor accounts for some of the variability in the motor commands that initiate saccades.

DESCRIPTION (provided by applicant): Our ability to adapt to systematic perturbations makes it possible to maintain a lifetime of calibrated movements. Our focus here is on the neural basis of this motor memory. The current view is that adaptation depends critically on the cerebellum. However, over the last two years we, and others, have made a series of observations that challenge this view of adaptation. Here we suggest a different view of the problem of motor adaptation based on the core hypothesis that the cerebellum is embedded in a larger network that includes motor cortical areas, and that more than one mechanism is involved in forming a motor memory. Specifically, we suggest that motor memory is a result of interaction of distinct components: one component associates motor commands with sensory consequences, resulting in a forward model; one component searches the motor space for output that can produce a rewarding outcome, resulting in exploration; a third component relies on repetition to associate the sensory feedback with the motor commands, resulting in a feedback- dependent controller. In Aim 1, we will test that idea that the function of M1 during adaptation is to encode a component of motor memory that depends on reinforced repetition of motor commands. In Aim 2, we will test the hypothesis that damage to the cerebellum affects only one component of motor memory, the ability to form memories that depend on sensory prediction errors (forward models), but spares the ability to learn from repetition of motor commands. Our projects are clinically important because if we are right in that there are multiple neural pathways to formation of motor memory, then damage to one component may benefit from rehabilitation procedures that focus on remaining healthy neural structures. Our projects are important from a basic science standpoint because: (1) our experiments can connect the cerebellar-centric field of adaptation which has focused on error-dependent learning, with cerebral cortex-centric field of motor learning which has focused on repetition-dependent processes; (2) our experiments have the power to explain what is being 'prepared' by the brain during the preparatory period before movement onset; and finally (3) our experiments have the potential to actually test computational ideas that are very much in fashion in the field of optima control, and ask whether they have any relevance to the neural basis of motor control.

The brain does not need the cerebellum to make movements. Rather, it needs the cerebellum to make accurate movements. The cerebellum endows the organism with the ability to internally monitor and correct ongoing motor commands. This monitoring requires the cerebellum to be able to predict errors that are about to happen, and correct for them before they occur. That is, the brain relies on the cerebellum to have accurate internal models that learn to predict sensory consequences of motor commands. However, it has been difficult to decipher how the cerebellum represents internal models: for many forms of behavior, including saccadic eye movements. By examining the relationship between simple spikes of Purkinje cells (P-cells) and behavior, this project will advance our understanding of computations in the cerebellum. Results of these studies could provide new avenues of rehabilitation for patients with cerebellar damage.

The new idea in this proposal is that the basic unit of computation in the cerebellum may not be a single P-cell or a randomly selected population of P-cells, but rather a specific group of P-cells wherein all the P-cells share the same preference for prediction error. Using this idea, the collaborative team of investigators from Johns Hopkins University and University of Washington has found that during saccades, the simple spikes of P-cells predict with exquisite accuracy future behavior of the eyes. The aim of this project is to understand how the cerebellum learns to make such accurate predictions. The investigators present a new paradigm, one in which sensory errors are perpendicular to the direction of motion of the eyes. This paradigm is interesting because behavior shows considerable richness: motor commands that arrive early in the movement appear to change little following error, but those that come late show both high learning and rapid forgetting. Using a combination of experiments and computational modelling, the investigators will test ideas that behavioral changes are due to the neural changes in the P-cells: micro-clusters that do not prefer the error express their learning in the acceleration phase of the movement, whereas those that prefer the error express their learning in the deceleration phase.