Wavelet-based Correlations of Skin Temperature and Blood Flow Oscillations

Abstract

The wavelet transform-based correlation analysis has been used to study skin temperature fluctuations caused by periodic changes in blood flow resulting from oscillations in vasomotor smooth muscle tone. We considered two cases, one in which temperature measurements and blood flow recordings by laser Doppler flowmetry are made simultaneously and another in which two temperature signals are measured concurrently. Twelve healthy subjects participated in our study. The gapped wavelet technique was used to suppress artifacts caused by boundary effects. Simultaneous recordings of skin temperature fluctuations and the signal of the laser Doppler flowmeter provided correlation coefficients essentially exceeding the values obtained for noise signals within three spectral ranges corresponding to myogenic (0.05–0.14 Hz), neurogenic (0.02–0.05 Hz), and endothelial (0.0095–0.02 Hz) regulation mechanisms. Within the frequency range from 0.14 to 2 Hz the values of the correlation function are compatible with the values of noise correlations. The same results were obtained for two concurrently measured temperature signals. Reduction in the amplitude of temperature fluctuations and in the level of correlations with the frequency arises because the skin has the properties of a low-frequency filter. As temperature fluctuations propagate their amplitude decays as an exponential function of frequency. Hence small oscillations in the spectral range reflecting the influence of heartbeat and respiration cannot be distinguished from external thermal noise.