Table 1. Results of the studies assessing (ultra)-short-term pulse rate variability from video recordings.
Results of the studies by Sun et al. (2012) and Cerina, Iozzia & Mainardi (2017) are shown for the maximum frame rate only, that is, 200 and 60 fps, respectively. ICC, intraclass correlation coefficient; IR, infrared; LRC, linear regression coefficient; MAE, mean absolute error; n, normalized values; r, correlation coefficient; r2, coefficient of determination; RMSE, root mean square error; SD, standard deviation; SD1, Poincaré plot standard deviation perpendicular to the line of identity; SD2, Poincaré plot standard deviation along the line of identity; SSE, sum of squared errors; VS, visible spectrum.
Study | Parameter | SDNN (ms) | RMSSD (ms) | pNN50 (%) | VLF (ms2 or n.u.) | LF (ms2 or n.u.) | HF (ms2 or n.u.) | LF/HF (%) | SD1 (ms) | SD2 (ms) |
---|---|---|---|---|---|---|---|---|---|---|
Poh, McDuff & Picard (2011) | Mean error | – | – | – | – | 7.53 (n) | 7.53 (n) | 0.57 | – | – |
SD of error | – | – | – | – | 10.17 (n) | 10.17 (n) | 0.98 | – | – | |
UST-PRV | RMSE | – | – | – | – | 12.3 (n) | 12.3 (n) | 1.1 | – | – |
r | – | – | – | – | 0.92 (n) | 0.92 (n) | 0.88 | – | – | |
Sun et al. (2012) ST-PRV |
r | 0.874* | – | – | – | 0.971* (n) | 0.978* (n) | 0.875* | – | – |
McDuff, Gontarek & Picard (2014) | r | – | – | – | – | 0.93** (n) | 0.93** (n) | 0.93** | – | – |
UST-PRV | RMSE | – | – | – | – | – | – | 0.145 | – | – |
Moreno et al. (2015) | r (supine and sitting) | 0.9544 | 0.8398 | 0.8635 | 0.9975/0.8928 (n) | 0.9849/0.7934 (n) | 0.9844/0.7838 (n) | 0.8662 | 0.8398 | 0.9734 |
0.9108 | 0.518 | 0.538 | 0.9979/0.8928 (n) | 0.9885/0.7934 (n) | 0.99/0.7838 (n) | 0.3186 | 0.518 | 0.9858 | ||
ST-PRV | Bias (supine and sitting) | 4.12 | 11.5 | 8.66 | 21.73/−2.18 (n) | 7.49/−0.53 (n) | 73.26/2.72 (n) | −0.07 | 8.13 | 2.47 |
6.68 | 19.3 | 16.87 | 4.96/−6.65 (n) | 13.69/−2.56 (n) | 206.83/9.21 (n) | −0.19 | 13.65 | 3.12 | ||
Effect size (supine and sitting) | 0.524 | 0.707 | 0.643 | 0.222/0.050 (n) | 0.006/0.050 (n) | 0.145/0.185 (n) | 0.226 | 0.707 | 0.244 | |
0.675 | 0.699 | 0.684 | 0.105/0.522 (n) | 0.280/0.285 (n) | 0.627/0.468 (n) | 0.257 | 0.669 | 0.631 | ||
Blackford, Piasecki & Estepp (2016) ST-PRV |
r2 | 0.9212 | – | – | – | 0.9728 | 0.7789 | – | – | – |
SSE | 7.3 | – | – | – | 0.18 | 0.38 | – | – | – | |
Bias | 12 | – | – | – | 0.13 | 0.78 | – | – | – | |
Huang & Dung (2016) UST-PRV |
MAE | 0.49–3.79 | 0.09–7.97 | – | – | – | – | – | 0.03–5.73 | 0.38–5.34 |
2.08–28.87 | 10.29–50.04 | – | – | – | – | – | 3.44–36.03 | 0.36–28.38 | ||
Iozzia, Cerina & Mainardi (2016) ST-PRV |
r (rest and stand) | 0.95 | 0.92 | 0.90 | – | 0.92 (n) | 0.91 (n) | 0.93 | – | – |
0.82 | 0.75 | 0.68 | – | 0.63 (n) | 0.76 (n) | 0.71 | – | – | ||
Purtov et al. (2016) ST-PRV |
r | – | – | – | >0.9 | >0.9 | – | – | – | – |
Cerina, Iozzia & Mainardi (2017) ST-PRV |
r | 0.9532 | 0.9433 | 0.9124 | – | 0.8368 (n) | 0.8372 (n) | – | – | – |
Melchor Rodriguez & Ramos-Castro (2018) UST-PRV |
r (stationary and motion) | 0.9750* | 0.9791* | 0.9350* | – | 0.9824* (n) | 0.9824* (n) | 0.9521* (n) | – | – |
0.9896* | 0.9463* | 0.9204* | – | 0.9288* (n) | 0.9302* (n) | 0.8548* (n) | – | – | ||
ICC (stationary and motion) | 0.9491* | 0.9208* | 0.8996* | – | 0.9546* (n) | 0.9546* (n) | 0.7593* (n) | – | – | |
0.9766* | 0.8898* | 0.8560* | – | 0.9247* (n) | 0.9265* (n) | 0.4468*** (n) | – | – | ||
MAE (stationary and motion) | 3.49 | 7.13 | 8.65 | – | 4.78 (n) | 4.77 (n) | 0.31 (n) | – | – | |
2.65 | 7.81 | 7.80 | – | 5.54 (n) | 5.47 (n) | 1.10 (n) | – | – | ||
RMSE (stationary and motion) | 4.26 | 8.29 | 10.35 | – | 6.20 (n) | 6.20 (n) | 0.61 (n) | – | – | |
3.34 | 9.34 | 10.30 | – | 7.44 (n) | 7.34 (n) | 3.28 (n) | – | – | ||
Yu et al. (2018) ST-PRV |
LRC | 0.99 (VS) | – | – | – | 1.09 (VS) | 1.12 (VS) | 1.10 (VS) | – | – |
0.95 (IR) | – | – | – | 1.11 (IR) | 1.07 (IR) | 1.09 (IR) | – | – | ||
Benezeth et al. (2018) UST-PRV |
Accuracy | – | – | – | – | 98.25% | 97.59% | 96.90% | – | – |
Notes:
p < 0.001.
p < 0.01.
p < 0.05.