Electrocorticography signals for human hand prosthetics

University Of Washington, Seattle WA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Neurological injury (such as from stroke, traumatic brain injury, and spinal cord injury) is a major cause of permanent disability. Recent advances in the field of neuroprosthetics hold enormous potential for the development of brain-computer interfaces to restore neurological function. This project will lead to a system that can control a robotic hand using recordings from the surface of the brain. Interfaces based directly from brain signals may allow for direct decoding of control signals for maximally efficient prosthetics. This project, a collaboration between neurosurgery, computer science, and physics departments, will explore the brain signals underlying hand movement using electrocorticography, or ECoG. We have previously shown that high frequency (>75Hz) components of the ECoG carry information about local brain activity. In the first aim, we will expand our understanding of the high-frequency signal components that correlate with individual finger movements. We will extract broadband changes in ECoG from non-specific alpha and beta rhythms using PCA and enhance finger classification with machine learning algorithms. In the second aim, we will look for control signals reflecting different hand functions, rather than movement of different fingers. For instance, we will examine if pinch and grasp behaviors give more separable high- frequency ECoG signals. We will also examine the behavior of these movements at higher spatial resolution. In the third aim, we will measure ECoG changes associated with imagined movement and how these changes are altered with visual feedback when applied to a robotic hand. In the final aim, we will add tactile feedback to the control to optimize ECoG-based control of a hand prosthesis. By increasingly advancing the complexity of the control signal, and the complexity of the robotic hand output, we will establish if ECoG is a viable source of control signal for a hand neuroprosthetic device.