Heart Rate Variability from Wearable Photoplethysmography Systems: Implications in Sleep Studies at High Altitude

Figures

• 

  Figure A1. HRV multiscale entropy (MSE) for m = 2 sleeping at high altitude. Upper panels: mean (solid line) ± standard error (dotted line) for MSE of DDI, SSI, dP1, and dP2 (black) and the RRI reference (red) on N = 21 participants (panels (a–d)). Lower panels: MSE ratio between DDI (e), SSI (f), dP1 (g), and dP2 (h) tachograms and the RRI reference: median (solid line) with 1st and 3rd quartiles (dotted line) and V-statistics of the Wilcoxon’s rank test; at the scales where V is above the 5% (or 1%) significance threshold (dashed horizontal line) the ratio differs from 1 at the 5% (or 1%) statistical significance. Grey/white areas indicate the scales for evaluating MSEHF, MSELF and MSEVLF.

• 

  Figure 1. ECG and PPG tachograms. Derivative-and-threshold algorithms are applied to (a) ECG, (b) intensity of transmitted light (TL) on the finger PPG sensor, (c) absorbed light (AL = −TL), (d) AL 1st and (e) 2nd derivative. RRI is the reference; PPG tachograms are: DDI = interval between TL maxima; SSI = interval between AL maxima; dP1 = interval between maxima of AL 1st derivative; and dP2 = interval between maxima of AL 2nd derivative.

• 

  Figure 2. Example of IBI series during sleep at high altitude. (a) RRI from the ECG (reference); (b) DDI from the transmitted light (TL); (c) SSI from the absorbed light (AL = −TL); (d) dP1 from AL 1st derivative and (e) dP2 from AL 2nd derivative.

• 

  Figure 3. HRV power spectra density (PSD) sleeping at high altitude. Upper panels: geometric mean (solid line) ± geometric standard error (dotted line) for PSD calculated from DDI (a), SSI (b), dP1 (c), and dP2 (d) tachograms (black) and the reference RRI (red) over N = 21 participants. Lower panels: spectral ratios between DDI (e), SSI (f), dP1 (g), and dP2 (h) tachograms and the reference tachogram: median (solid line) with 1st and 3rd quartiles (dotted line) and V-statistics of Wilcoxon rank test; at frequencies where the V-value is above the 5% (or 1%) significance threshold (dashed horizontal line) the spectral ratio differs from 1 at the 5% (or 1%) statistical significance. Grey/white areas indicate the frequency bands for evaluating VLF, LF, and HF powers.

• 

  Figure 4. HRV detrended fluctuation analysis (DFA) sleeping at high altitude. Upper panels: mean (solid line) ± standard error (dotted line) for DFA spectra of DDI (a), SSI (b), dP1 (c), and dP2 (d) (black) and the reference RRI (red) on n = 21 participants. Lower panels: ratio between the DFA of DDI (e), SSI (f), dP1 (g), and dP2 (h) tachograms and the reference RRI: median (solid line) with 1st and 3rd quartiles (dotted line) and V-statistics of Wilcoxon rank test; at scales where the V-value is above the 5% (or 1%) significance threshold (dashed horizontal lines) the DFA ratio differs from 1 at the 5% (or 1%) statistical significance. Grey/white areas indicate the scales for evaluating α1 and α2 scale exponents.

• 

  Figure 5. HRV multiscale entropy (MSE) for m = 1 sleeping at high altitude. Upper panels: mean (solid line) ± standard error (dotted line) for MSE of DDI, SSI, dP1, and dP2 (black) and the RRI reference (red) on n = 21 participants (panels (a–d)). Lower panels: MSE ratio between DDI (e), SSI (f), dP1 (g), and dP2 (h) tachograms and the RRI reference: median (solid line) with 1st and 3rd quartiles (dotted line) and V-statistics of the Wilcoxon’s rank test; at the scales where V is above the 5% (or 1%) significance threshold (dashed horizontal line) the ratio differs from 1 at the 5% (or 1%) statistical significance. Grey/white areas indicate the scales for evaluating MSEHF, MSELF and MSEVLF.