Wavelet analysis of autonomic and cardiovascular signals

Abstract

The autonomic nervous system is a regulatory structure primarily responsible for the physiologic response to environmental change, which includes the regulation of heart rate and blood pressure. Accurate assessment of autonomic function is important in the study and diagnosis of various disorders. In the research setting, autonomic sympathetic function can be assessed by directly recording the sympathetic nerve activity whereas, clinically, indirect autonomic assessment using heart rate and blood pressure variability is more practical. In this work, wavelet-based signal processing is used to improve the existing quantification schemes for both direct and indirect autonomic function. An action potential detection scheme involving the stationary wavelet transform was developed as a direct quantifier of human sympathetic nerve activity. The detection performance of this method was validated using simulated signals with varying burst rate and signal to noise ratio. Sympathetic spike detection also demonstrated a strong correlation to common integrated burst parameters in data acquired during graded upright tilt. A similar wavelet-based spike detection method was also designed for the mouse sympathetic nerve activity in order to identify differences in sympathetic function in transgenic models of autonomic disorders. A simple model was then created which used the sympathetic spike rate and the instantaneous respiration to explain the neurally-mediated low frequency oscillations and the mechanically-driven high frequency fluctuations observed in the continuous blood pressure. The model revealed a strong correlation between the low frequency fluctuations in blood pressure and spike rate at rest and during orthostatic stress. This result suggested that the low frequency blood pressure patterns could be used as an indirect estimate of sympathetic activity. Finally, a form of time varying spectral analysis capable of quantifying the power of the low and high frequency oscillations in non-stationary heart rate and blood pressure data was developed using a modified wavelet transform. This method was used to identify potential differences in the autonomic mechanisms associated with syncope alone and syncope ending in asystole, a rare stoppage of the heart, in a normal population during orthostatic stress. In summary, wavelet-based techniques are useful tools to study autonomic-cardiovascular regulation.