Higher-order spectral analysis in laser-Doppler flowmetry signal processing.

This paper presents an application of higher order statistics and spectra to the problem of laser-Doppler Flowmetry (LDF) for microcirculation monitoring. The proposed signal processing technique computes the bispectra of LDF signals after appropriate pre-processing and extracts features which can be used for classification of the signals to a normal or patient category. Bispectra are defined in terms of the third-order moments or cumulants of signals and are shown to contain valuable information for the above classification of LDF signals. Experimental studies, including (a) a set of 17 normal subjects and 69 patients with Raynaud's phenomenon and (b) 50 LDF-signals separated to four groups from patients that have different pathologic characteristics, are described and results are presented which illustrate the performance of the proposed approach when applied to the LDF signals.