Investigating the Capabilities of Novel Silk Sericin-based Electrodes to Measure Electrocardiogram Signals by Using Machine Learning Techniques

In this paper, novel compostable non-invasive and self-adhesive silk sericin-based electrodes, designed for recording 1-lead ECG signals, have been investigated through machine learning approaches. ECG signals, recorded through two different types of silk sericin-based innovative electrodes, with a varying percentage of calcium chloride (20 and 30 wt%), were assessed by comparing their performance against each other and against commercial Ag/AgCl electrodes. The comparison was carried out by considering both relevant temporal intervals of the ECG signals themselves and Heart Rate Variability (HRV) features processed through different machine learning algorithms. Additionally, we also studied the ability of innovative sericin-based electrodes to detect stress and fatigue states, obtaining significant results upon comparison with commercial electrodes. The silk sericin-based electrodes we tested can thus be suitable for effective on-the-move ECG monitoring and possible integration into future smart healthcare systems.