Causal Network in a Deafferented Non-Human Primate Brain

Karthikeyan Balasubramanian, Kazutaka Takahashi, Nicholas Hatsopoulos

Details

Category

Invited session papers

Theme

06. Neural and Rehabilitation Engineering

Sessions

08:30 - 10:00 | Wed 26 Aug | Brown 3 | 6.1

Brain-Computer/Machine Interface I

Abstract

De-afferented/efferented neural ensembles can undergo causal changes when interfaced to neuroprosthetic devices. These changes occur via recruitment or isolation of neurons, alterations in functional connectivity within the ensemble and/or changes in the role of neurons, i.e., excitatory/inhibitory. In this work, emergence of a causal network and changes in the dynamics are demonstrated for a deafferented brain region exposed to BMI (brain-machine interface) learning. The BMI was controlling a robot for reach-and-grasp behavior. And, the motor cortical regions used for the BMI were deafferented due to chronic amputation, and ensembles of neurons were decoded for velocity control of the multi-DOF robot. A generalized linear model-framework based Granger causality (GLM-GC) technique was used in estimating the ensemble connectivity. Model selection was based on the AIC (Akaike Information Criterion).

Additional Information

No information added

Video

No videos found