Development of an EMG Based Tic Detector

Objective: To develop a surface electromyography (EMG) technique to detect and characterize motor tics.

Background: Tourette syndrome is a complex neuropsychiatric disorder characterized by a combination of motor and vocal tics. The clinical evaluation of tics is complex because of difficulties relating to the variability, suppressibility, and suggestibility of tics as well as issues with the limitations of available evaluation tools. Currently, available tools rely on an individual subject’s recollection of tics or alternatively on a short potentially non-representative in-clinic video. These limitations can have implications in practice and in clinical trials.

Design/Methods: Wireless EMG sensors (Trigno by Delsys^™) were applied over the extremity and trunk muscles of subjects with Tourette syndrome. The EMG was rectified and referenced to baseline. Spectral analysis used 0.5 Hz bins. EMG data features most predictive of tics were inputted into a support vector machine. The output of the support vector machine was a binary decision of “Tic” or “No Tic”. The EMG data were used to classify the tics by body parts involved, frequency, and severity.

Results: Power spectral analysis of the movement epochs was used to calculate EMG signal power. Data were recorded from multiple muscles during voluntary movements and tic movements. Differences were analyzed for statistical significance. This is an ongoing study and we will present the results of the cohort at the AAN meeting in May 2019.

Conclusions: Preliminary analysis reveals that EMG features could be a promising tool to differentiate tics from voluntary movement.