On Improving the Effectiveness of Control Signals from Chronic Microelectrodes for Cortical Neuroprostheses.

Abstract

Using microelectrodes, we can record neural signals which can eventually be used to control cortical neuroprostheses for assisting people with spinal-cord trauma, stroke deficits, amyotrophic lateral sclerosis (ALS), and motor-neuron disease. The goal of this dissertation is to investigate the effectiveness of unit activity and local field potentials (LFPs) in the motor cortex using chronic multisite microelectrodes.

In the first study, we first demonstrate a novel method to assess neural signatures across sessions and quantify neuron stability by providing a probabilistic estimate of similarity between spike clusters. This technique supports both single and multiple electrodes, and has applications in designing appropriate neuroprosthetic control algorithms, determining recalibration parameters, investigating neural plasticity, and assessing significance of particular metrics.

Next, we investigate unit activity and LFP activity in the different layers of the motor cortex. Four rats were implanted bilaterally with multi-site single-shank silicon microelectrode arrays in the motor cortex while the animal was engaged in a movement-direction task. In the second study, we demonstrate that units in the lower layers (Layers 5,6) are more likely to encode direction information as compared to units in the upper layers (Layers 2,3) suggesting electrode sites clustered in the lower layers provide access to more salient control information.

In the third study, we investigate LFP activity across the different layers. We analyzed LFP activity in four frequency ranges: low (3-15Hz), low-gamma (15-40Hz), high-gamma (40-70Hz) and high (>70Hz) across both upper (Layers 2,3) and lower layers (Layers 5,6) of the cortex. Our analysis based on 585 LFP recordings from 39 sessions shows that the low frequency range (3-15Hz) is more likely to encode directional information. We found a significant difference in LFP activity between the upper and lower layers of cortex in the high gamma (40-70Hz) range. Our results indicate that LFPs are viable alternative control signals that can be recorded from either upper or lower layers of the cortex for performance comparable to our results from unit activity.