Robert A. Scheidt, Ph.D. - US grants

0238442
Scheidt
Engineers have long drawn inspiration and insight from the study of biological systems. Some of the most intriguing parallels between biological and man-made systems have arisen in the study of information processing and control within biological nervous systems. Early explorations into control and communication in animals and machines have influenced the development of diverse engineering disciplines such as artificial intelligence, computational neuroscience and the mathematical modeling of logic and the mind. Particularly intriguing are studies of learning and adaptation in nervous systems; such studies have influenced the development and analysis of artificial neural networks and have in turn been influenced by advances in adaptive control theory.

The goals of the proposed research are to identify the neural mechanisms mediating adaptive improvements in motor performance, and to test the hypothesis that abnormal sensory feedback gain degrades motor coordination in a relatively common clinical population. The PI will develop a novel, single-degree-of-freedom, robotic manipulandum for use with both event-related functional magnetic resonance imaging (ER-fMRI) and systems identification techniques to characterize, model and locate within the brain the adaptive mechanisms mediating an important form of motor learning known as motor adaptation. Neurologically unimpaired subjects and autistic children will make goal-directed hand movements while holding the handle of a novel robot. ER-fMRI techniques will be used to quantify brain activity while the robot perturbs the moving hand. Linear (and if needed, nonlinear) systems identification techniques will be used to identify which brain region(s) are most likely involved in the adaptive neural representation of the hand's mechanical environment.

The proposal outlines the PI's role in developing Marquette University's new Biocomputer Engineering curriculum. The PI will train up to 40 engineers per year in the processes, methodologies, technologies and physiological aspects of engineering microcontroller-based systems in the regulated medical electronics industry. The PI will integrate state-of-the-art knowledge of adaptive algorithms and their validation into the new undergraduate curriculum. The PI will also integrate the proposed research and device development activities into his existing Neuromotor Control course. Finally, the proposal describes an engineering curriculum development process that draws from industrial best practices in product design and quality control. Curriculum development will incorporate industrial advisory feedback from periodic curriculum design reviews to ensure that state-of-the-art design processes, methodologies and technologies are instructed.

[unreadable] DESCRIPTION (provided by applicant): Project Summary/Abstract This project analyzes disordered control of posture and movement during reaching in patients with stroke resulting from middle cerebral artery occlusion. Prior studies have shown that control of posture and movement may both be impaired, and contribute importantly to disability. Our recent work has shown that normal reaching requires control of both posture and movement, that these two functions are specified by different neural mechanisms, and that accuracy is dependent on adaptation (learning) and coordination between the two. Based on our preliminary observations, we hypothesize that stroke degrades the ability to recruit and relax the balanced muscle co-contractions needed both to overcome limb postural bias associated with hypertonia and to stabilize the hand against unsteady loads across the workspace. We further hypothesize that major impairment in reaching post-stroke arises from improper scaling and timing of muscle coactivation, thus limiting the independence of arm trajectory and position control. [unreadable] Using a planar robot arm and novel EMG biofeedback methods, Aim 1 proposes to characterize the stroke- related changes in: 1) the ability to maintain postural stability throughout the workspace via graded coactivation of antagonist muscles; 2) the latency and developmental time course of different levels of coactivation, and 3) the use of proprioceptive feedback to counteract reflex abnormalities during regulation of limb position in the presence of mechanical perturbations. [unreadable] Aim 2 examines stroke-related changes in the integration of posture and movement control. Using the planar robot we will: 1) explore transfer of learning between posture and movement tasks, assessing whether coupling between the two controllers increases post-stroke, and 2) determine whether trajectory planning adapts post-stroke to account for the biomechanical effects of posture regulation. [unreadable] We expect our results will ultimately lead to new approaches for rehabilitating arm function post-stroke. We anticipate that improvement in arm function will be promoted by applying robotic and biofeedback methods first to train coactivation, posture and movement control separately, then in combination. [unreadable] [unreadable]

Current research and clinical efforts related to post-stroke arm rehabilitation focus primarily on motor retraining, with limited focus on the impact of somatosensory deficits on motor function. This is not surprising given that arms are not very useful without volitional movement. However, somatosensory deficits are common in the contralesional arm and may contribute importantly to deficits in the control of functional movement. This project advances the goal of promoting functional motor recovery after stroke by creating sensory substitution technologies that re-establish kinesthetic feedback control of the contralesional arm by delivering augmented feedback to a body part for which the brain retains the ability to process somatosensory feedback. The objective of this application is to determine how best to synthesize and deliver supplemental kinesthetic feedback, and to test its ability to enhance sensorimotor control over the contralesional arm post-stroke. This study has two Aims. The first seeks to optimize delivery of supplemental kinesthetic feedback to enhance reach, stabilization and manipulation actions of the contralesional arm post-stroke. Several different vibrotactile feedback encodings of limb position and velocity will be synthesized and applied to sites on the body retaining somatosensation. The specific combination of state variables and stimulation site that best enhance stabilization and manipulation with the contralesional arm and hand will be identified in a small cohort of stroke survivors. This Aim tests the hypotheses that supplemental feedback including both position and velocity state information will best enhance arm control and best reduce abnormal coupling between hand grip force and arm stability. Aim 2, seeks to characterize learning that accrues due to extended training with supplemental kinesthetic feedback. Over a period of three weeks, a small cohort of stroke survivors will train to use supplemental kinesthetic feedback to enhance reach-to-grasp actions in a 3-dimensional environment. We test the hypothesis that extended training with supplemental kinesthetic feedback leads to new compensatory skills that generalize to untrained action sequences that contribute to the success of many behaviors of daily living. Upon completion, this project will determine how best to synthesize and deliver supplemental sensory feedback to improve contralesional arm control in stroke survivors with residual motor capacity but impaired or absent proprioception in the contralesional arm. This contribution will be significant because it develops a new assistive technology with the potential to improve contralesional arm use in many stroke survivors. This proposal is innovative because it represents a substantive departure from the status quo, both in the field of physical rehabilitation after stroke and in the field of sensory substitution, by shifting the focus of motor retraining toward the re-establishment of real-time closed-loop feedback control of the contralesional arm. Successful completion of this project will ultimately lead to novel, continuously wearable technologies that will enable many stroke survivors to recover impaired or lost capabilities in the contralesional arm.

Project Summary Sensorimotor impairment is common after stroke. Sensorimotor impairment can degrade quality of life by limiting simple acts like reaching to catch a child's hand. This project is based on the premise that the individualization of therapeutic intervention needed to optimize motor function after neuromotor injury is predicated on an understanding how healthy brains integrate sensory, motor, and cognitive functioning to control movement. Our project focuses on reaching to moving objects using a cued manual interception task because interception underlies many activities of daily life in an unpredictable environment. People and objects do not always move as expected, making mid-movement error corrections critical for success. Our project will advance a fundamental understanding of how the brain adjusts ongoing actions to correct performance errors that arise mid-movement. We focus on neurologically intact people as a foundational step toward our long-term goal, which is to develop and test intervention strategies to address the impact of each patient's neural injury on spatiotemporal and dynamic aspects of error correction in everyday activities. Cued manual interception affords ideal experimental control over the timing of events giving rise to movement and error. Our innovative project exploits the precision motion and measurement abilities of a rehabilitation robot, the fine temporal resolution of electroencephalography (EEG), and a highly multidisciplinary team to advance understanding of the behavioral and neural basis of mid-movement error correction during manual interception. We will characterize behavioral and neural responses to three different sources of error: inherent errors in the selection and execution of action, unpredictable target motion, and unpredictable environmental dynamics resisting hand motion. Error correction is significant to study because it enables success in everyday tasks, particularly when things do not go as initially planned. We will conduct human subjects experiments that will address two specific aims. Neurologically-intact people will hold the handle of a planar robot while trying to catch a moving visual target. Infrequently, at reach onset target speed will increase or the robot will render an unexpected change to the hand's load. Analysis of behavioral data (Aim 1) will determine the dynamics of behavioral corrections in response to target interception errors. We hypothesize that response latencies reflect differences in information processing required to correct errors arising from the three different error sources. Analysis of EEG data (Aim 2) will identify patterns of neural activity and functional connectivity leading to error corrections in response to target interception errors. We hypothesize that neural correlates of error correction reflect differences in the type and timing of information processing required to compensate for the three sources of error. If successful, this R21 project will lead to new knowledge about the spatiotemporal patterns of neural activity in response to errors that arise when reaching toward moving targets. This project will also identify neural circuits contributing to the initiation of movement, to the on-line detection of performance errors, and to the generation of corrective actions in response to those errors. Advancing understanding of mid-movement error correction will facilitate the development of therapeutic interventions designed to enhance functional movement and quality of life in patients with neuromotor injury.