Abstract:Cardiorespiratory fitness has been shown to be correlated with cardiovascular complications and comparable to other conventional health parameters. Consequently, it can be a useful preventative and a practical first line of information for cardiovascular diseases, such as hypertension. Wearable devices have the capacity to provide a convenient assessment of individuals cardiorespiratory fitness. Despite its clinical relevance measures of cardiorespiratory fitness remain limited to commercial devices for exercise training purposes. This study investigates whether there is potential for cardiorespiratory fitness to be derived from indices measured at rest from a finger photoplethysmogram waveform. Participants (n = 8, 4 females, 39 ± 17 years) first completed two questionnaires to provide two separate estimates of cardiorespiratory fitness. The first questionnaire by Jackson et al. (1990) provides an estimated maximum rate of oxygen (VO2max, gold standard measure of cardiorespiratory fitness) based on height, weight, age, and a subjective physical activity rating score. The second questionnaire by Radim et al. (2005) provides a metabolic equivalent parameter (MET) that has previously been shown to correlate with measured VO2max. Following 10 minutes of seated rest, 30 seconds of finger photoplethysmography was recorded. A 10-second snippet with high quality signal was extracted, and a generalised transfer function was applied to the signal to generate an uncalibrated central aortic waveform. Pulse pressure amplification (PPAmp), ejection duration (ED), subendocardial viability ratio (SEVR), heart rate (HR) and augmentation index corrected to a HR of 75 bpm (AIx75bpm) were then calculated for each cardiac cycle and averaged over 10 seconds. Correlation plots and stepwise linear regression were then performed on the estimated VO2max and MET with the derived parameters. Average baseline estimated VO2max was 44 ± 10 ml/kg/min and estimated MET was 11 ± 2 (1 MET = VO2/3.5). Stepwise linear regression revealed AIx75bpm, SEVR and HR were the strongest predictors of estimated VO2max (model R^2 = 0.929, F = 17.44, p = 0.009). Only AIx75bpm and SEVR were strongly associated with estimated MET (model R^2 = 0.942, F = 40.5, p < 0.001). For both estimated VO2max and estimated MET, individual correlation plots for PPAmp and ED showed poor correlation. A selection of non-exercise derived central aortic parameters from finger photoplethysmography were correlated with cardiorespiratory fitness. This demonstrates that a finger photoplethysmogram signal obtained at rest, such as can be easily measured in wearable devices, may be useful in management and monitoring of cardiorespiratory fitness, and by association health, without the burden of performing rigorous exercise tests.

Toward characterizing cardiovascular fitness using machine learning based on unobtrusive data

Longitudinal cardio-respiratory fitness prediction through wearables in free-living environments

Nonexercise machine learning models for maximal oxygen uptake prediction in national population surveys

Predicting physical fitness levels from resting ECG data: a machine learning approach in cardiovascular assessment

Assessing cardiorespiratory fitness in clinical and community settings: Lessons and advancements in the 100th year anniversary of VO2max

Effect of urban environment on cardiovascular health: a feasibility pilot study using machine learning to predict heart rate variability in patients with heart failure

PS-C25-5: ASSOCIATIONS BETWEEN CARDIORESPIRATORY FITNESS AND CARDIOVASCULAR PARAMETERS DERIVED AT REST FROM FINGER PHOTOPLETHYSMOGRAM

Association of Smartwatch-Based Heart Rate and Physical Activity With Cardiorespiratory Fitness Measures in the Community: Cohort Study

Heart rate prediction with contactless active assisted living technology: a smart home approach for older adults

Am I fit for this physical activity? Neural embedding of physical conditioning from inertial sensors

Machine Learning Models for the Automatic Detection of Exercise Thresholds in Cardiopulmonary Exercising Tests: From Regression to Generation to Explanation

Improving cardiovascular risk stratification through multivariate time-series analysis of cardiopulmonary exercise test data

The Development and Validation of a New Tool for Spanish Adults

Predicting Soccer Players' Fitness Status Through a Machine-Learning Approach

Population and Age-Based Cardiorespiratory Fitness Level Investigation and Automatic Prediction

Machine Learning on Prediction of Relative Physical Activity Intensity Using Medical Radar Sensor and 3D Accelerometer

Learning Individualized Cardiovascular Responses from Large-scale Wearable Sensors Data

Prediction of peak oxygen consumption using cardiorespiratory parameters from warmup and submaximal stage of treadmill cardiopulmonary exercise test

Unsupervised Clustering of Heartbeat Dynamics Allows for Real Time and Personalized Improvement in Cardiovascular Fitness

Predicting the occurrence of minor adverse events in cardiac rehabilitation using physical variables

Quality Assessment of Photoplethysmography Signals For Cardiovascular Biomarkers Monitoring Using Wearable Devices