Blepharospasm (BSP) is characterized by stereotyped, bilateral, and synchronous spasm of the orbicularis oculi muscles. Additional motor manifestations include apraxia of eyelid opening and increased blink rate. Current rating scales for BSP are subjective, based on rater observation, and therefore limited in their ability to measure these motor manifestations and guide treatment. Accurate measurement of BSP symptoms is critical for optimization of clinical care for this particularly disabling condition and for future clinical trials.

Subjects with a diagnosis of idiopathic BSP were recruited. The severity of BSP was assessed using the current validated clinical rating scales. Objective data quantifying dystonic movements were obtained using a set of soft, stretchable, gel-free electrodes which were placed over the frontalis and a portion of the orbicularis oculi to measure movements. Portable, miniaturized wireless electronics were utilized to record the signals via an Android smartphone. To analyze the real-time recorded data, we used a deep-learning method (convolutional neural network) that was embedded in the smartphone. The newly developed software offered an automatic, simultaneous detection of the motor manifestations of BSP along with a digital scale severity.