Electro-Stimulation System with AI-based auricular-triggered Algorithm for Patients with Peripheral Facial Palsy

Abstract

Facial Palsy causes severe functional disorders and impairs quality of life. Two of the most disturbing challenges for patients with facial palsy are the loss of the ability to smile and insu cient eyelid closure. A potential treatment of these conditions could be a closed-loop electro-stimulation system that stimulates the facial muscles on the paretic side as needed to elicit eye closure, eye blink and smile similar to the healthy side. This study focuses on developing such a system. A ML algorithm classi es the intended facial movements based on the auricular muscle EMGs of the paretic side. The system then delivers targeted surface stimulation to the appropriate facial muscles based on this classi cation. The present work details the development of the system and the integrated AI-based auricular-triggered algorithm. The evaluation of both was conducted with 17 patients. The study demonstrated that auricular EMG signals of the paretic side contain su cient information for use in a closed-loop electro-stimulation system. The algorithm achieved a median macro F1-score of 0.708. Additionally, the study showed that such a system, using an AIbased auricular-triggered algorithm, can support to some level facial movements on the paretic side in patients with peripheral facial palsy. The system itself reached a median macro F1-score of 0.602.

Date of Award	May 2024
Original language	English
Supervisor	Thomas Haslwanter (Supervisor)