Issue 1/2 Vol. 37 1995

Kalman Filter Algorithm in the investigation of physiological oscillations

SCHMITZ, N., KUGLER, J. & NEUMANN, W.

Heart rate variabilities convey important information regarding central nervous system regulation. Spectral analysis is a powerful instrument in frequency domain to estimate and analyse periodicities. A basic statistical assumption of spectral analysis is that the mean and variance of a heart rate signal do not change with time. This assumption is often untrue in psychophsiological experiments. An alternative approach is to use a recursive spectral estimation with a Kalman Filter Algorithm to estimate time dependent frequencies. This method is superior to ordinary spectral analysis for time series with changing frequencies.

Keywords: central nervous system regulation, spectral analysis, Kalman Filter Algorithm

Short Title: Schmitz, N. et al. (1995) PsyBeit 1-2:11

[Pabst Science Publishers] [Psychologische Beiträge] [Table of Contents] [Search] [Order]