Abstract Details
| Presented By: | Kellis, Spencer |
| Affiliated with: | University of Utah, Electrical Engineering |
| Authors: | Spencer Kellis, Paul House, Kyle E Thomson, Richard Brown, Bradley Greger |
| From: | University of Utah |
Title
Abstract
The most successful approaches to correlating neuronal activity with movement used penetrating electrodes to record the firing patterns of multiple single neurons in primates. However, penetrating electrodes have in vivo lifetimes of only a few years, with loss of signal quality over time. Additionally, chronically penetrating electrodes have difficulty maintaining recordings from an individual neurons, likely because of micromotion of the device relative to the neural tissue. ECoG and EEG nonpenetrating electrodes have been used successfully in neuroprosthetic applications, but have low spatial resolution due to large electrodes and spacing. We investigated a relatively new platform, nonpenetrating microwire arrays, which have the potential for high spatial fidelity and good chronic performance.
Two male patients were implanted with nonpenetrating microwire arrays (40-micron wire; 1-2mm spacing). Both performed repetitive motor tasks involving movement of the hand and arm contralateral to the hemisphere in which the arrays were implanted. An analysis of velocity profiles was used to extract 1-second windows of neural data, aligned to movement at 0.5 seconds.
Channel-pairwise cross-correlation performed on this data demonstrated decreasing correlation with distance. This result suggests the nonpeneterating microwire array captures more sources of information within the area covered by a single macroelectrode.
Trial-averaged spectrograms revealed patterning consistent with previous analyses, including modulation in the beta band (12-30Hz) power before and during movement. Additionally, the percent percent change in gamma band power (30-80Hz), during planning phase, between movement in the contralateral and ipsilateral directions, was reasonably substantial, suggesting binary classification of these directions is possible using millimeter-spaced microwires.
This work demonstrates that nonpenetrating mirowire arrays record neocortical activity with improved spatial resolution than larger surface electrodes. Additionally, neural data recorded from nonpenetrating microwires supported discrimination of gross direction based on modulation in gamma band power. By providing high spatial resolution recordings of human neocortical activity, nonpenetrating microwire arrays may be able to serve as an interface that provides intuitive control of a neuroprosthesis, and may serve as a novel research tool for studying spatially local neocortical phenomena in health and disease.