Face-wearable integrated bioelectronics for quantitative, automated diagnosis of blepharospasm

Abstract

Blepharospasm (BSP) is a neuro-ophthalmologic disorder marked by excessive blinking and involuntary contractions of the muscles around the eyes. Current standard clinical evaluations rely mainly on subjective assessments, often resulting in inconsistencies and human errors in diagnosis and severity monitoring. Here, we introduce a wireless, face-wearable, all-in-one bioelectronic system designed to continuously capture high-fidelity electrooculograms and electromyograms as a quantitative tool for diagnosing BSP. This device features soft membrane sensors and integrated circuits that ensure skin conformity, allowing for highly accurate signal detection on the face. The wearable system has been optimized in both design and functionality to detect a wide range of BSP-related issues across multiple patients, such as increased blink rates, eyelid fluttering, and prolonged eye closures. Our study shows that the normal blink frequency is similar (p = 0.546); however, the BSP group exhibits longer durations and higher amplitudes (p < 0.005). Partial blinks are more frequent and have higher amplitudes, but similar durations (p < 0.005). Long blinks are different in both frequency and duration, but not amplitude (p < 0.01). Flutter events also show group differences in frequency (p < 0.01) and duration (p < 0.005), with no amplitude difference (p = 0.168). A machine learning-based prediction model demonstrates an accuracy of 81.5 % and an F1-score of 0.814 when validated against expert-annotated video data. Overall, the combination of wireless soft bioelectronics and advanced machine learning algorithms, presented in this work, shows a first-of-a-kind approach to effectively and accurately diagnosing BSP.