Deji Akinwande1,Dmitry Kireev1,Kaan Sel2,Bassem Ibrahim2,Roozbeh Jafari2
The University of Texas at Austin1,Texas A&M University2
Deji Akinwande1,Dmitry Kireev1,Kaan Sel2,Bassem Ibrahim2,Roozbeh Jafari2
The University of Texas at Austin1,Texas A&M University2
Monitoring complex health-related electrophysiological signals such as arterial blood pressure (BP) in ambulatory settings is essential for a proper understanding of health conditions, predominantly cardiovascular diseases. Moreover, continuous long-term monitoring of BP for patients with sleep apnea, stroke, or hypertension is essential to understand their health risk factors and build preventative care routines. While conventional ambulatory BP monitoring devices exist, they are uncomfortable, bulky, and intrusive. The common drawbacks of all these systems are their bulkiness and incompatibility with skin’s elastic properties, causing sensor’s displacement during usage, consequently requiring frequent system re-calibration.<br/>Here, we present a unique wearable BP monitoring platform that leverages imperceptible atomically-thin and electrically conductive graphene electronic tattoos (GETs) as the main building block. The GETs are placed over the radial and ulnar arteries on the wrist and subsequently used as current injection and voltage sensing electrodes, measuring arterial bioimpedance. In contrast to any other wearable system, the atomically thin, lightweight, and skin-conformable GETs do not apply any external tension onto the skin during the operation, and able to perform long-term and nocturnal measurements without discomforting the subjects. Using bioimpedance modality allows us to disregard the tattoo-skin interface, which is typically 2-4 orders of magnitude larger compared to tissue impedance, and record only from the areas of interest. Employing a machine learning regression model on the recorded bioimpedance value, we yield effective beat-to-beat detection of diastolic and systolic BP values with IEEE grade-A accuracy. Besides BP, we show that the same Bio-Z signal can be post-processed to estimate person’s RR in an entirely wearable and non-invasive manner.