AI-enabled epidermal electronic system to automatically monitor a prognostic parameter for hypertension with a smartphone

Abstract

We present a wearable, flexible, wireless and smartphone-enabled epidermal electronic system (EES) for the continuous monitoring of a prognostic parameter for hypertension. The thin and lightweight EES can be tightly attached to the chest of a patient and synchronously monitor first lead electrocardiograms (ECG) and seismocardiograms (SCG). To demonstrate the concept, we developed the EES using state-of-the-art cleanroom technologies. Two types of sensors were integrated: A pair of metal electrodes to contact the skin and to record ECG and a vibration sensor based on a thin piezoelectric polymer to record SCG from the same location of the chest, simultaneously. The complete EES was powered by the near field communication functionality of the smartphone. We developed a machine-learning algorithm and trained it on public ECG data and recorded SCG signals to extract characteristic features of the recordings. Binary classifiers were used to automatically annotate peaks. After training, the algorithm was transferred to the smartphone to continuously analyze the timing between particular ECG and SCG peaks and to extract the Weissler's index as a prognostic parameter for hypertension. Tests with data of healthy control persons and clinical experiments with patients diagnosed with cardio-pulmonary hypertension showed a promising prognostic performance. The presented EES technology could be utilized for pre-screening of cardio-pulmonary hypertension, which is a strong burden in our today's healthcare system.

Keywords: AI-based data analysis; Automatic signal classification; Cardiac monitoring; Cardio-pulmonary hypertension; Sensor fusion; Wearable sensors.