Beyond Blood Pressure and Heart Rate Monitoring: Towards a Device for Continuous Sensing and Automatic Feature Extraction of Cardiovascular Data

Abstract

Heart disease contributes significantly to annual deaths. Personal health monitoring systems can be useful in reducing mortality rates as well as for treatment and prevention of associated complications. A current challenge is the limited accessibility to devices which offer advanced feature extraction for cardiovascular diagnosis. In this work, investigations are presented using non-invasive and continuous monitoring using photoplethysmography (PPG), a simple, low-cost, popular choice for wearable devices, from which features other than pulse rate and blood pressure can be extracted from acquired PPG signals. The investigations outlined in this paper advance work in automated feature extraction for cardiovascular diagnosis, demonstrated through identification of the 'a', 'b' and 'e' waves derived from the second derivative of the PPG waveform, followed by calculation of indices associated with patient arterial stiffness using these waves. Results suggest that the implemented approach is a useful tool in assessing physiological features which could be utilized to determine cardiovascular conditions. The approach is important in providing insights into the emerging design space for cardiovascular monitoring and feature extraction as part of personalized healthcare systems.