Search

This paper presents a review of our neural interface research program and provides a brief introduction to the field. We focus on three key areas which aim to minimize encapsulating tissue response to chronic neural interfaces: interface architecture, mechanical stresses, and stem cell-seeded probes. We explore these areas and present our current solutions which have led to the demonstration of cortical control of neuromotor prostheses (NMPs).

While the development of silicon-substrate microelectrode arrays has enabled chronic recording of single unit activity from multiple neurons simultaneously, accurate interpretation of the signals depend on the anatomical placement of the electrodes. Toward this end, this paper develops an in vivo method for identifying the placement of electrodes based on laminar analysis of ketamine-xylazien-induced field potential oscillations in rat motor cortex. The proposed method is based on finding the polarity reversal in laminar oscillations which is reported to appear in upper part of layer V in laminar High Voltage Spindles (HVSs) of rat cortical column. Analysis of histological images showed a 21 mum error in the estimate of the polarity reversal depth compared to the expected range (850-1050 mum). One out of the four rats did not undergo a phase reversal. Histology verified that the electrode was placed deeper than 1050 jim. We propose that this method can be used to determine an estimate of laminar electrodes implanted in rat motor cortex

Developments of the past decades have shown it is possible to decode and extract control signals directly from the brain for use in neuromotor prosthetics. Currently, there is a push towards developing closed loop interface systems. Here we describe our early results in a rat model of a sensory/motor neurophysiological feedback control system. Traditional operant conditioning techniques historically train an animal to increase or decrease the occurrence of a behavior in response to a stimulus using various types of reinforcement. Here we demonstrate a rat performing an operant conditioning task with the modification that the rat's behavioral output and stimulus input are accessed via direct electrical connections with the brain. Our preliminary results here show a rat successfully trained to modulate the firing rates of its motor cortex in response to intra-cortical microstimulation of the visual cortex.

The interface between neural prostheses and neural tissue plays a significant role in the long term performance of these devices. Conducting polymers have been used to modify the electrical properties of neural microelectrodes. The objective of this study was to evaluate recording chronic neural activity of neural microelectrodes that were modified with nanofibers-templated of poly (3,4-ethylenedioxythiophene) (PEDOT) nanotubes over seven week periods using impedance spectroscopy and signal-to-noise ratio measurements. PEDOT nanotubes-coated sites were found to have lower impedance and higher signal-to-noise ratio than control site.

Brain machine interface development typically falls into two arenas, invasive extracellular recording and non-invasive electroencephalogram recording methods. The relationship between action potentials and field potentials is not well understood, and investigation of interrelationships may improve design of neuroprosthetic control systems. Rats were trained on a motor learning task whereby they had to insert their noses into an aperture while simultaneously pressing down on levers with their forepaws; spikes, local field potentials (LFPs), and electrocorticograms (ECoGs) over the motor cortex were recorded and characterized. Preliminary results suggest that the LFP activity in lower cortical layers oscillates with the ECoG.

Traditional brain-machine interfaces have typically focused on methods that use rate-based codes as a source for control signals. Opposed to rate, timing of firing across different neurons and within each neuron could also provide information that can be used for controlling brain-machine interfaces or neuroprosthetic devices. Findings have indicated that synchronization of individual spike discharges may help serve the organization of cortical motor processes. We are investigating neural firing synchrony in the context of using it for real-time control for neuroprostheses systems. Our results with rats suggest that subjects can be trained to synchronize neural firing and increase unitary events i.e. spike coincidence patterns that are significantly above chance. Temporal coding methods could be used as additional or alternative cortical control signals for neuroprostheses and brain machine interfaces.