Central and peripheral autonomic influences : analysis of cardio-pulmonary dynamics using novel wavelet statistical methods

CENTRAL AND PERIPHERAL AUTONOMIC INFLUENCES: ANALYSIS OF CARDIOPULMONARY DYNAMICS USING NOVEL WAVELET STATISTICAL METHODS by Anne Marie Petrock The development and implementation of novel signal processing techniques, particularly with regard to applications in the clinical environment, is critical to bringing computeraided diagnoses of disease to reality. One of the most confounding factors in the field of cardiac autonomic response (CAR) research is the influence of the coupling of respiratory oscillations with cardiac oscillations. This research had three objectives. The first was the assessment of central autonomic influence over heart rate oscillations when the pulmonary system is damaged. The second was to assess the link between peripheral and central autonomic control schema by evaluating the heart rate variability (HRV) of people who were able or unable to adapt to the use of integrated lenses for vision, specifically accommodation, correction (adaptive and non-adaptive presbyopes). The third objective was the development of a wavelet-based toolset by which the first two objectives could be achieved. The first tool is a wavelet based entropy measure that quantifies the level of information by assessing not only the entropy levels, but also the distribution of the entropy across frequency bands. The second tool is a wavelet source separation (WavS) method used to separate the respiratory component from the cardiac component, thereby allowing for analysis of the dynamics of the cardiac signal without the confounding influence of the respiratory signal that occurs when the body is perturbed. With regard to hypothesis one, the entropy method was used to separate the COPD study populations with 93% classification accuracy at rest, and with 100% accuracy during exercise. Changes in COPD and control autonomic markers were evident after respiration is removed. Specifically, the LF/HF ratio slightly decreased on average from pre to post reconstruction for controls, increased on average for COPD. In healthy controls, respiration frequency is distributed across multiple bandwidths, causing large decreases in both LF and HF when removed. With respiration effect removed from COPD population, LF dominates autonomic response, indicating that the frequency is concentrated in the HF autonomic region. Decrease in variance of data set increases probability that smaller changes can be detected in values. The theory set forth in hypothesis two was validated by the quantification of a correlation between peripheral and central autonomic influences, as evidenced by differences in oculomotor adaptability correlating with differences in HRV. Standard Deviation varies with grouping, not with age. Increasing controlled respiration frequencies resulted in adaptive presbyopes and controls displaying similar sympathetic responses, diverging from non-adaptive group. WayS reduced frequency content in ranges concurrent with breathing rate, indicating a robust analysis. The outcome of hypothesis three was the confirmation that wavelet statistical methods possess significant potential for applications in HRV. Entropy can be used in conjunction with cluster analysis to classify patient populations with high accuracy. Using the WayS analysis, the respiration effect can be removed from HRV data sets, providing new insights into autonomic alterations, both central and peripheral, in disease, CENTRAL AND PERIPHERAL AUTONOMIC INFLUENCES: ANALYSIS OF CARDlO-PULMONARY DYNAMICS USING NOVEL WAVELET STATISTICAL METHODS by Anne Marie Petrock A Dissertation Submitted to the Faculty of New Jersey Institute of Technology in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Biomedical Engineering Department of Biomedical Engineering