Attempt to Evaluate Autonomic Nervous System Activity on Intake of Functional Foods Using Empirical Mode Decomposition: A Comparison with Conventional Methods

Background and objectives: The effects of functional foods on autonomic nervous system activity (ANSA) have been investigated via heart rate variability (HRV). HRV is the variability of heartbeat intervals and includes linear and nonlinear components. In conventional frequency analysis of HRV, the fast Fourier transform (FFT) has been used for linear analysis. Additionally, a combination of FIR filtering and the Hilbert transform (HT) has been used for nonlinear analysis in some studies. Recently, a combination of empirical mode decomposition (EMD) and the Hilbert transform has been proposed as more suitable for nonlinear signals. In this study, the FFT, HT, and EMD analyses were performed on HRV before and after functional food intake, and the results were compared. Methods: Six healthy young males participated in two experimental sessions with gamma-aminobutyric acid (GABA) or placebo intake. Electrocardiograms were recorded for six minutes each five times before and after intake. A mental arithmetic stress task was applied for 17 minutes starting 13 minutes after intake. HRV data were obtained from the electrocardiograms and three analysis methods were performed. The low frequency band (LF: 0.04–0.15 Hz) component, high frequency band (HF: 0.15–0.4 Hz) component and LF/HF ratio were calculated as indices of ANSA via HRV. Results: In the EMD analysis, HF and LF/HF ratio were not obtained for only one subject. LF/HF ratio of the subject in the FFT and HT analyses were two to three times higher than that of the other subjects. In another subject, the standard deviation of normal to normal (SDNN), a time-domain index of HRV, was outside the healthy range. LF and HF of the subject were up to 30 times higher than that of the remaining four subjects in the FFT analysis, but up to four times higher in the HT and EMD analyses. In the remaining four subjects, the HT and EMD analyses tended to show less inter-subject variability in all HRV indices compared to the FFT analysis. Conclusions: We conclude that the HT and EMD analyses offer the possibility of examining parameter changes in HRV in more detail on intake of functional foods.