Statistical selection of multiple-input multiple-output nonlinear dynamic models of spike train transformation.

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)22_Publication in policy or professional journal

View graph of relations

Author(s)

  • Dong Song
  • Vasilis Z Marmarelis
  • Robert E Hampson
  • Sam A Deadwyler
  • Theodore W Berger

Detail(s)

Original languageEnglish
Pages (from-to)4727-4730
Journal / PublicationConference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference
Publication statusPublished - 2007
Externally publishedYes

Abstract

Multiple-input multiple-output nonlinear dynamic model of spike train to spike train transformations was previously formulated for hippocampal-cortical prostheses. This paper further described the statistical methods of selecting significant inputs (self-terms) and interactions between inputs (cross-terms) of this Volterra kernel-based model. In our approach, model structure was determined by progressively adding self-terms and cross-terms using a forward stepwise model selection technique. Model coefficients were then pruned based on Wald test. Results showed that the reduced kernel models, which contained much fewer coefficients than the full Volterra kernel model, gave good fits to the novel data. These models could be used to analyze the functional interactions between neurons during behavior.

Citation Format(s)

Statistical selection of multiple-input multiple-output nonlinear dynamic models of spike train transformation. / Song, Dong; Chan, Rosa H M; Marmarelis, Vasilis Z; Hampson, Robert E; Deadwyler, Sam A; Berger, Theodore W.

In: Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Conference, 2007, p. 4727-4730.

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)22_Publication in policy or professional journal