Robust carotid pulse detection using accelerometry and electrocardiography

Abstract

Neurally mediated syncope (NMS) is a disorder of the autonomic regulation of postural tone. These patients suffer from blood-pressure regulation failure resulting in hypo-perfusion of the brain, which might lead to significant injuries due to falls. Detection of the onset of blood-pressure regulation failure enables the patient to start counter-measures and, hence, avoid hypo-perfusion of the brain. Monitoring of blood-pressure (BP) changes is key to develop an early alert solution. Several setups have been tested for this purpose using vascular unloading and pulse arrival time (PAT). The aim of this study was to do an initial investigation of using the electrocardiography (ECG) signal and accelerometry (ACC) signals acquired from the carotid to reliably infer pulse presence and pulse absence. A classifier using a correlation feature derived from the ECG and the ACC signals was conceptualized and implemented. The algorithm performance was tested with data acquired during scheduled head-up tilt table tests of 8 syncope patients. This is an essential step towards PAT inference as a BP surrogate.