A Proof-of-Concept Study on Noncontact BCG-Based Cardiac Monitoring for In-Patients With Sleep Apnea Syndrome Using Piezoelectric Ceramics

Cardiac monitoring is essential for in-patients with sleep apnea syndrome (SAS), where ballistocardiography (BCG)-based techniques provide a contact-free alternative. To assess the potential of BCG for cardiac monitoring in practical clinic scenario, this study conducted consistency analysis between vital signs simultaneously recorded by BCG and electrocardiogram (ECG), where the latter has been regarded as the gold standard. To be specific, 37 SAS in-patients of various severities from the First Affiliate Hospital of Guangzhou Medical University were recruited for whole-night BCG and ECG recording using piezoelectric ceramics and Alice 5 polysomnography (PSG, manufactured by Philips Respironics Company), respectively. The off-line BCG-based signal processing was then undertaken, including artifact and signal quality (SQ) labeling, heartbeat interval detection, as well as heart rate variability (HRV) evaluation. The numerical results showed that the intrinsic heartbeat intervals derived from BCG and ECG signals were significantly consistent, where the overall average absolute error was less than 14 ms. The accuracy of detected heartbeat intervals was not only determined by the detection method, but also traded off against the coverage. With the aid of SQ labeling, we found that the consistence of heartbeat interval detection by different methods can be underlying indicator for BCG SQ assessment. Besides, the fusion of heartbeat interval detection methods complement each other and improve the average precision by up to 12.84% which demonstrated the potential of adaptive heartbeat interval detection in practical scenarios. Specially, mild and well-healed patients with SAS enjoyed both higher heartbeat interval detection precision and higher coverage, which implied the great potential of noncontact BCG-based cardiac monitoring for prescreening and follow-up of SAS patients.