The Assessment of Autonomic Nervous System Activity Based on Photoplethysmography in Healthy Young Men

Abstract

Noninvasive assessment of autonomic nervous system (ANS) activity is of great importance, but the accuracy of the method used, which is primarily based on electrocardiogram-derived heart rate variability (HRV), has long been suspected. We investigated the feasibility of photoplethysmography (PPG) in ANS evaluation. Data of 32 healthy young men under four different ANS activation patterns were recorded: baseline, slow deep breathing (parasympathetic activation), cold pressor test (peripheral sympathetic activation), and mental arithmetic test (cardiac sympathetic activation). We extracted 110 PPG-based features to construct classification models for the four ANS activation patterns. Using interpretable models based on random forest, the main PPG features related to ANS activation were obtained. Results showed that pulse rate variability (PRV) exhibited similar changes to HRV across the different experiments. The four ANS patterns could be better classified using more PPG-based features compared with using HRV or PRV features, for which the classification accuracies were 0.80, 0.56, and 0.57, respectively. Sensitive features of parasympathetic activation included features of nonlinear (sample entropy), frequency, and time domains of PRV. Sensitive features of sympathetic activation were features of the amplitude and frequency domain of PRV of the PPG derivatives. Subsequently, these sensitive PPG-based features were used to fit the improved HRV parameters. The fitting results were acceptable (p < 0.01), which might provide a better method of evaluating ANS activity using PPG.

Keywords: autonomic nervous system; cardiovascular system; classification; heart rate variability; noninvasive assessment technique; photoplethysmography.

Figure 1

Feature extraction of PPG and its derivatives. (A) Feature points on PPG, VPG (first derivative or velocity of PPG), and APG (second derivative or acceleration of PPG). (B) Representative features of the PPG. (C) Representative features of the PPG. O: onset of the systolic wave; S: the systolic peak; M: the midpoint of the systolic peak; N: the dicrotic notch; D: the diastolic peak; w: the maximum slope point in systole; z: the maximum slope point in diastole. a: the maximum acceleration point in systole; b: the maximum negative acceleration point in the falling edge; e: the maximum acceleration point in diastole. PW, pulse width corresponding to the midpoint of the rising phase; PPT, peak to peak interval between the systolic and diastolic peaks; TW, tidal wave; TW_t, time range of TW; TW_v, amplitude of TW; TW_a, area of TW.

Figure 2

Heart rate variability (HRV) and pulse rate variability (PRV) for different ANS activation patterns. (A) LF, low-frequency power. (B) HF, high-frequency power. (C) nLF, normalized low-frequency power. (D) LF/HF, ratio of LF to HF. BSL, baseline state; SDB, slow deep breathing; CPT, cold pressor test; MAT, mental arithmetic test. The significance of the differences in HRV parameters between each experimental state and BSL state is denoted by * for p < 0.05 and ** for p < 0.01. ^▴ and ^▴▴ indicate p < 0.05 and p < 0.01, respectively, for the corresponding comparisons of the PRV parameters. The physiological mechanism of HRV parameters is considered as follows: LF mainly contains information on sympathetic activity, HF reflects parasympathetic activity, nLF is a relatively pure measure of sympathetic activity, and LF/HF reflects sympathovagal balance.

Figure 3

The classification accuracies of the models based on heart rate variability (HRV), pulse rate variability (PRV), and all PPG-based features (mean accuracy of 30 runs at different random states).

Figure 4

The distribution of interpretation features in different ANS states. For each feature, the data were normalized between the four states. Red (baseline state, BSL), blue (slow deep breathing, SDB), black (cold pressor test, CPT), and pink (mental arithmetic test, MAT). The diamond mark represents the mean value. Std, standard deviations; Rmssd, root mean square of the successive difference; H1, height of the systolic peak; RS, rising slope of PPG; T, time span of the pulse; T3, diastolic time of the pulse; TW_st, stress index; Hw, height of w point in VPG; Hz, height of z point in VPG; Ha, height of a point in APG; Hb, height of b point in APG; He, height of e point in APG; SampEn, sample entropy of pulse rate; LF, low-frequency power of PRV; TP, total power of PRV. nLF, normalized low-frequency power; LF/HF, ratio of LF to HF.

Figure 5

The fitting results of the improved HRV parameters using the PPG features and stepwise regression. (A) LFru, low-frequency power of the improved HRV. (B) HFr, high-frequency power of the improved HRV. BSL, baseline state; SDB, slow deep breathing; CPT, cold pressor test; MAT, mental arithmetic test. The physiological mechanism of the improved HRV parameters are considered as: LFru mainly contains information on sympathetic activity, HFr reflects parasympathetic activity.

Figure 6

The statistical results of slow deep breathing (SDB), the cold pressor test (CPT), and the mental arithmetic test (MAT) compared with the baseline state (BSL) for the improved HRV parameters and fitting data. (A) LFru, low-frequency power of the improved HRV. (B) HFr, high-frequency power of the improved HRV. (C) nLFru, normalized low-frequency power. (D) LFru/HFr, the ratio of LFru to HFr. The significance of the differences in the improved HRV parameters between each experimental state and BSL state is denoted by * for p < 0.05 and ** for p < 0.01. ▴ and ▴▴ indicate p < 0.05 and p < 0.01, respectively, for the corresponding comparisons of the fitted result. The physiological mechanism of the improved HRV parameters are considered as follows: LFru mainly contains information on sympathetic activity, HFr reflects parasympathetic activity, nLFru is a relatively pure measure of sympathetic activity, and LFru/HFr reflects sympathovagal balance.