HEART RATE VARIABILITY measured using facial video photoplethysmography demonstrated good accuracy when compared with electrocardiogram (ECG) reference measurements in healthy adults, according to a pilot validation study.
Researchers evaluated whether a smartphone camera could estimate two commonly used heart rate variability indices, standard deviation of normal heartbeat time intervals (SDNN) and natural logarithm of the root mean square of successive differences in time intervals between heartbeats (lnRMSSD), from one minute facial video recordings.
The study included 35 healthy volunteers, including 17 females, with a median age of 25 years.
Participants remained seated and kept their heads relatively still while researchers recorded one minute facial videos using a smartphone mounted on a tripod.
Simultaneously, a one lead ECG was recorded to provide reference values for heart rate variability.
Video photoplethysmography is a contactless technique that detects subtle changes in skin colour caused by blood vessel pulsations.
These signals can be analysed to estimate physiological parameters, including heart rate variability.
The investigators aimed to determine how closely facial video derived measurements matched established ECG based assessments.
Analysis revealed strong agreement between facial video and ECG derived heart rate variability values.
For SDNN, the median absolute error was 3.0 ms, equivalent to 7.1%, while the root mean square error was 4.5 ms or 15.7%.
For lnRMSSD, the median absolute error was 0.20, corresponding to 5.6%, and the root mean square error was 0.31 or 9.9%.
Correlations between the two methods were strong. The Pearson correlation coefficient reached 0.98 for SDNN and 0.88 for lnRMSSD, with p<0.001 for both measures.
Researchers reported that 34 of 35 SDNN estimates were within 10 ms of the ECG reference values, while nearly 90% of lnRMSSD estimates were within 0.5 of the corresponding ECG measurement.
Although the technology generally performed well, investigators observed a slight tendency to overestimate heart rate variability values, particularly when measurements were low. This bias was evident for both SDNN and lnRMSSD.
The researchers concluded that facial video photoplethysmography can provide accurate estimates of heart rate variability in young healthy adults under controlled conditions.
However, they emphasised that further research is needed in larger populations with greater diversity in age and skin tone to confirm the findings and determine broader clinical applicability.
Reference
Pstras L et al. Estimating heart rate variability using facial video photoplethysmography: a pilot validation study. Sci Rep. 2026;DOI:10.1038/s41598-026-70292-0.
Featured image: Kicking Studio on Adobe Stock
Each article is made available under the terms of the Creative Commons Attribution-Non Commercial 4.0 License.
