(Georgia Institute of Technology, 2022-07-29)
Johnson, John Terry
The human ability to perform powerful, complex, and intricate actions using our hands is made possible by the hand's 27 degrees of freedom, its rich sensory feedback, and our ability to observe and understand the actions of others. In the event of an upper-extremity amputation, not only are articulated biological structures lost, but also their available sensory feedback. While great advances have been made in prosthesis engineering, the rate of prosthesis use in upper-extremity amputees remains low. Those who choose to use a prosthesis only do so an average of 50% of the time they could use them. One proposed strategy to improve prosthesis use and acceptance is the provision of task-salient vibrotactile feedback. Studies have been conducted on the efficacy of vibrotactile feedback, but have largely focused on kinematics, leaving a dearth of knowledge regarding the neural effects of vibrotactile feedback. In addition to changes in carrying out reach-to-grasp tasks, amputation and prosthesis use result in a mismatch between the amputee's end-effector (usually a split-hook), and that of non-amputated persons using their hand. This mismatch may lead to increased difficulty determining the actions of others. Sensorimotor models relating motor actions and their resulting sensory feedback also provide the basis for understanding the intentions of other people as we observe their actions. Brain areas which generate our own movements based on our goals and intentions also relate movements we see performed by others to their probable motor commands. When observing movements, the kinematics we see have probable motor commands, which in turn lead to probable goals, and goals to the probable intention of the action. Thus the lack of sensorimotor models for using a prosthesis can lead to not only difficulty using the prosthesis, but also determining the actions of others. The purpose of the proposed studies is to evaluate the neural and behavioral effects of vibrotactile feedback during prosthesis use, and to evaluate cognitive changes between observing a hand or a prosthesis performing reach-to-grasp actions. In Aim 1, electroencephalography and 3D motion capture were used to examine neural activity while performing reach-to-grasp actions using a prosthesis with and without vibrotactile feedback. In Aim2, functional magnetic resonance imaging was used to measure differences in brain activity when naïve participants unfamiliar with prostheses observed a hand and a prosthesis reaching to grasp everyday objects. In Aim 3, clusters of cortical activity found in Aim 2 were used to assess effective connectivity between clusters, as well as to determine graph-theoretic network properties of those clusters, and their changes when observing either a hand or prosthesis performing reach-to-grasp actions. Overall, findings from these studies reveal anticipation of upcoming vibrotactile feedback in sensorimotor brain areas. When observing actions performed with a prosthesis, findings show a diminished recruitment of brain areas specialized for determining the intention of others when observing their actions. The neural activity anticipating upcoming somatosensory feedback, as well as the changes in neural activity when observing a prosthesis rather than a hand, provide insight into the cognitive processes during prosthesis use, and increases our understanding of the role of sensory feedback in those processes.