• DocumentCode
    1397996
  • Title

    Bispectrum analysis of visually evoked potentials

  • Author

    Husar, P. ; Henning, G.

  • Author_Institution
    Fac. for Inf. & Autom., Tech. Hochschule Ilmenau, Germany
  • Volume
    16
  • Issue
    1
  • fYear
    1997
  • Firstpage
    57
  • Lastpage
    63
  • Abstract
    This article presents results obtained from single-set analysis, and an increase in the SNR is derived and demonstrated by real data. In addition, some aspects of practical signal analysis and detection are discussed. It is concluded that, being a relatively simple method for analyzing biosignals, the bispectrum turns out to be very efficient for the detection of phase-coupled harmonics in background noise. The results obtained so far for steady-state visually evoked potentials suggest that there are nonlinearities of at least seventh order in the conduction and processing of the stimuli. Therefore, future research will be carried out to detect this phenomenon. A necessary condition for using conventional detectors is the whiteness of noise. Since higher-order spectra are used in the analysis of biosignals, an appropriate prewhitening algorithm for higher-order moments needs to be developed.
  • Keywords
    medical signal processing; spectral analysis; visual evoked potentials; SNR; background noise; biosignals analysis; bispectrum analysis; higher-order moments; noise whiteness; phase-coupled harmonics; prewhitening algorithm; seventh order nonlinearities; single-set analysis; steady-state visually evoked potentials; Biomedical engineering; Biomedical informatics; Biomedical signal processing; Frequency; Retina; Signal analysis; Signal processing; Signal to noise ratio; Tires; Visual system; Algorithms; Artifacts; Biomedical Engineering; Electroencephalography; Evoked Potentials, Visual; Humans; Models, Biological; Reaction Time; Reproducibility of Results; Signal Processing, Computer-Assisted; Visual Fields;
  • fLanguage
    English
  • Journal_Title
    Engineering in Medicine and Biology Magazine, IEEE
  • Publisher
    ieee
  • ISSN
    0739-5175
  • Type

    jour

  • DOI
    10.1109/51.566155
  • Filename
    566155