Divide-and-conquer approach for brain machine interfaces: nonlinear mixture of competitive linear models.

Published

Journal Article

This paper proposes a divide-and-conquer strategy for designing brain machine interfaces. A nonlinear combination of competitively trained local linear models (experts) is used to identify the mapping from neuronal activity in cortical areas associated with arm movement to the hand position of a primate. The proposed architecture and the training algorithm are described in detail and numerical performance comparisons with alternative linear and nonlinear modeling approaches, including time-delay neural networks and recursive multilayer perceptrons, are presented. This new strategy allows training the local linear models using normalized LMS and using a relatively smaller nonlinear network to efficiently combine the predictions of the linear experts. This leads to savings in computational requirements, while the performance is still similar to a large fully nonlinear network.

Full Text

Duke Authors

Cited Authors

  • Kim, S-P; Sanchez, JC; Erdogmus, D; Rao, YN; Wessberg, J; Principe, JC; Nicolelis, M

Published Date

  • June 2003

Published In

Volume / Issue

  • 16 / 5-6

Start / End Page

  • 865 - 871

PubMed ID

  • 12850045

Pubmed Central ID

  • 12850045

International Standard Serial Number (ISSN)

  • 0893-6080

Digital Object Identifier (DOI)

  • 10.1016/S0893-6080(03)00108-4

Language

  • eng

Conference Location

  • United States