Optimization and pre-use suitability selection for wrist photoplethysmography-based heart rate monitoring in patients with cardiac disease

Figures

• 

  Figure 1. Methodology overview of the three research parts aiming to evaluate and optimize the accuracy of wrist photoplethysmography based heart rate monitoring. CR, cardiac rehabilitation; ECG, electrocardiogram; HR, heart rate; MAPE, mean absolute percentage error; PPG, photoplethysmography. High accuracy is defined as MAPE <10% for ≥70% of the training time. Low accuracy is defined as MAPE <10% for <70% of the training time.

• 

  Figure 2. Accuracy metrics during various wrist movements with the PPG sensor placed one and three fingers from the wrist joint: (A) intraclass correlation coefficient, (B) MAE, (C) MAPE, and (D) total percentage of under- and overshooting. Bpm, beats per minute; ICC, intraclass correlation coefficient; MAE, mean absolute error; MAPE, mean absolute percentage error; PPG, photoplethysmography. Boxplots in panel B and C indicate the median, interquartile range (IQR), and whiskers extending to 1.5 × IQR; crosses represent the mean. Under/overshooting was defined as a percentage difference of 10% or more compared with the Polar device.

• 

  Figure 3. Accuracy metrics during various sport activities: (A) intraclass correlation coefficient (ICC), (B) MAE, (C) MAPE, and (D) total percentage of under- and overshooting. Bpm, beats per minute; ICC, intraclass correlation coefficient; MAE, mean absolute error; MAPE, mean absolute percentage error; PPG, photoplethysmography. Boxplots in panel B and C indicate the median, interquartile range (IQR), and whiskers extending to 1.5 × IQR; crosses represent the mean. Under/overshooting was defined as a percentage difference of 10% or more compared with the Polar device.

• 

  Figure 4. Percentage of accurate training time (MAPE <10%) per participant, per assessment moment, and split for arm movement intensity. MAPE, mean absolute percentage error. High accuracy is defined as MAPE <10% for ≥70% of training time (everything above the black threshold dotted line). Low accuracy is defined as MAPE <10% for <70% of training time (everything below the black threshold dotted line). Following exercises are classified as intensive arm movements: arm bike, rowing machine. Following exercises are classified as non-intensive arm movements: strength training (including leg press, chest press, low row, vertical traction, recumbent bike), cross-trainer, stairmaster, bike, walking/running on treadmill.

• 

  Figure 5. Percentage of accurate training time (MAPE <10%) by exercise type for baseline assessment (n = 30). MAPE: mean absolute percentage error. Results are ranked from the lowest to the highest accuracy.