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

Sung Phil Kim, Justin C. Sanchez, Deniz Erdogmus, Yadunandana N. Rao, Johan Wessberg, Jose C. Principe, Miguel Nicolelis

Research output: Contribution to journalArticlepeer-review

52 Scopus citations

Abstract

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.

Original languageEnglish (US)
Pages (from-to)865-871
Number of pages7
JournalNeural Networks
Volume16
Issue number5-6
DOIs
StatePublished - 2003

Keywords

  • Brain machine interfaces
  • Competitive learning
  • Multiple local linear models

ASJC Scopus subject areas

  • Artificial Intelligence
  • Neuroscience(all)

Fingerprint Dive into the research topics of 'Divide-and-conquer approach for brain machine interfaces: Nonlinear mixture of competitive linear models'. Together they form a unique fingerprint.

Cite this