Properties of a temporal difference reinforcement learning brain machine interface driven by a simulated motor cortex

Aditya Tarigoppula; Nick Rotella; Joseph T Francis

doi:10.1109/EMBC.2012.6346666

Properties of a temporal difference reinforcement learning brain machine interface driven by a simulated motor cortex

Annu Int Conf IEEE Eng Med Biol Soc. 2012:2012:3284-7. doi: 10.1109/EMBC.2012.6346666.

Authors

Aditya Tarigoppula¹, Nick Rotella, Joseph T Francis

Affiliation

¹ Department of Physiology and Pharmacology, SUNY Downstate Medical Center, Brooklyn, 450 Clarkson Avenue, Brooklyn, NY 11203, USA. aditya30887@gmail.com

PMID: 23366627
DOI: 10.1109/EMBC.2012.6346666

Abstract

Our overall goal is to develop a reinforcement learning (RL) based decoder for brain machine interfaces. As an important step in this process, we determine the basic stability and convergence properties of a Temporal Difference (TD) RL architecture being driven by a simulated motor cortex.

Publication types

Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

Brain-Computer Interfaces*
Learning*
Motor Cortex / cytology
Motor Cortex / physiology*
Neurons / physiology
Signal-To-Noise Ratio