Discrimination of cardiac health and disease by assessment of heart rate variability: wavelet vs. fast Fourier transformation

The autonomic nervous system (ANS) modulation of the heart is of clinical importance because of its relevance to risk of life threatening arrhythmic events. Decomposition of heart rate variability (HRV) has been used to quantify ANS control of the heart. The traditional method for frequency analysis has involved the use of fast Fourier transformation (FFT). However, heart rate data typically violate assumptions of the FFT. Therefore, the assessment of HRV may benefit from other, potentially more suitable, mathematical approaches. For example, the discrete wavelet transformation (DWT) appears to have promise with respect to its ability to discriminate between healthy and diseased populations. Therefore, the purpose of this thesis was to examine the extent to which the FFT can discriminate between a control group and heart failure patients (CHF) in comparison to DWT. Seven CHF (mean +/- standard deviation, age: 51.9 +/- 17.6 yrs) and eight age-matched controls (49.5 +/- 17.9 yrs) were evaluated. HRV was evaluated during 5 minutes of supine spontaneous breathing (SB) and supine paced breathing (PB) (0.2Hz). The ECG data were sampled at 200 Hz, converted to heart rate tachograms, and subjected to frequency analysis via FFT and DWT. The FFT approach did not reveal group differences in HRV, while the DWT revealed group differences in LF/HF during SB (p\u3c0.05) and PB (p=0.053). With respect to breathing condition, only the FFT revealed that PB resulted in a decrease in low- to high-frequency ratios (p\u3c0.05), and an increase in standard deviation of normal R-R intervals. These results support further consideration of both methods of analysis, as they each appear to provide unique information about HRV