Application of eigenstructure based bispectrum estimation: EEG wave coupling in cognitive tasks

David L. Sherman, Michael D. Zoltowski

Research output: Contribution to conferencePaperpeer-review

Abstract

A bispectral application is developed by considering a fully two-dimensional region of support in the cumulant domain. An asymptotically Hermitian block-Toeplitz matrix composed of third-order cumulants is built. Using a noise subspace-based frequency estimator, a two-dimensional frequency kernel is projected on the orthogonal complement subspace. The reciprocal of this projection is the frequency estimator indicating the location of a triple of phase-coupled sinusoids in the bispectral domain. The method provides accurate, high-order spectrum estimation. In general, transfer function parameterization is implicated in higher-order spectral estimation when one-dimensional or diagonal slices are used.

Original languageEnglish (US)
Pages135-139
Number of pages5
StatePublished - Dec 1 1989
EventWorkshop on Higher-Order Spectral Analysis - Vail, CO, USA
Duration: Jun 28 1989Jun 30 1989

Other

OtherWorkshop on Higher-Order Spectral Analysis
CityVail, CO, USA
Period6/28/896/30/89

ASJC Scopus subject areas

  • Engineering(all)

Fingerprint Dive into the research topics of 'Application of eigenstructure based bispectrum estimation: EEG wave coupling in cognitive tasks'. Together they form a unique fingerprint.

Cite this