Wavelet analysis of laser Doppler microcirculatory signals: Current applications and limitations

Abstract

Laser Doppler flowmetry (LDF) has long been considered a gold standard for non-invasive assessment of skin microvascular function. Due to the laser Doppler (LD) microcirculatory signal's complex biological and physiological context, using spectral analysis is advisable to extract as many of the signal's properties as feasible. Spectral analysis can be performed using either a classical Fourier transform (FT) technique, which has the disadvantage of not being able to localize a signal in time, or wavelet analysis (WA), which provides both the time and frequency localization of the inspected signal. So far, WA of LD microcirculatory signals has revealed five characteristic frequency intervals, ranging from 0.005 to 2 Hz, each of which being related to a specific physiological influence modulating skin microcirculatory response, providing for a more thorough analysis of the signals measured in healthy and diseased individuals. Even though WA is a valuable tool for analyzing and evaluating LDF-measured microcirculatory signals, limitations remain, resulting in a lack of analytical standardization. As a more accurate assessment of human skin microcirculation may better enhance the prognosis of diseases marked by microvascular dysfunction, searching for improvements to the WA method is crucial from the clinical point of view. Accordingly, we have summarized and discussed WA application and its limitations when evaluating LD microcirculatory signals, and presented insight into possible future improvements. We adopted a novel strategy when presenting the findings of recent studies using WA by focusing on frequency intervals to contrast the findings of the various studies undertaken thus far and highlight their disparities.

Keywords: endothelial function; frequency band; laser Doppler flowmetry; microcirculation; spectral analysis; wavelet analysis.

FIGURE 1

(A) Three-dimensional wavelet transform (WT) of the laser Doppler (LD) signal of a representative tracing (baseline period, finger pulp); and (B) Time-averaged wavelet power spectrum of the same signal and frequency intervals corresponding to endothelial nitric oxide (NO)-dependent (endo NO), neurogenic (neuro), myogenic (myo), respiratory (resp), and cardiac (card) physiological influences. The sample axis on the three-dimensional spectrum represents the sampling points of the signal. Thereby, it can also be observed as a time axis (A).