Resting and Post-exercise Heart Rate Detection from Fingertip and Facial Photoplethysmography Using a Smartphone Camera

Modern smartphones allow measurement of heart rate (HR) via detection of pulsatile photoplethysmographic (PPG) signals from the fingertips or the face without physical contact by extracting subtle beat-to-beat variations of skin color that is similar to HR fluctuations with built-in cameras. PURPOSE: To evaluate HR measurements at rest and after exercise using a smartphone based PPG detection application. METHODS: Forty healthy subjects (50% males; mean age 24.7 ± 5.2 years; Von Luschan skin color range 14-27) underwent treadmill exercise using the Bruce protocol. Simultaneous PPG signals were recorded per subject by (i) facing the front camera and (ii) placing index fingertip over the back camera of iPhone 6Ss and analyzed by the Cardiio-Heart Rate Monitor + 7 Minute Workout Exercise Routine for Cardio Health and Fitness (Cardiio Inc., Cambridge, MA) smartphone application for HR compared to continuous ECG as reference. Recordings of 20 seconds duration each were acquired at rest, immediately post moderate (50-70% maximum HR) and vigorous (70-85% maximum HR) intensity exercise and repeated successively until return to resting HR. Bland-Altman plots were used to examine agreement between ECG and PPG estimated HR. HR monitors are considered accurate if the correlation coefficient (r) is ≥0.93 and root mean square error (RMSE) is ≤5 beats per minute (bpm) or <6.8%. RESULTS: Mean HR was 73±13 bpm (range 53-114) at rest and 97±21 bpm (51 to 157) post exercise. 477 fingertip and 582 facial PPG measurements were recorded and analyzed. Fingertip PPG estimated HR were strongly correlated with both resting ECG HR (r = 0.997 and RMSE = 1.025 bpm or 1.404%) and post-exercise HR (r = 0.995 & RMSE = 2.074 bpm or 2.220%). Correlation of facial PPG estimated HR were stronger with resting ECG HR (r = 0.997 & RMSE = 1.020 bpm or 1.428%) and was also strong with post-exercise HR (r = 0.984 & RMSE = 3.673 bpm or 3.931%). Bland-Altman plots showed better agreement between ECG with fingertip than facial PPG estimated HR. CONCLUSIONS: HR detection by the Cardiio smartphone application is considered accurate at rest and after moderate and vigorous intensity exercise in a healthy young adult sample. Although touchless facial PPG detection is convenient, it may not be able to consistently detect a HR when body motion is excessive after exercise.