JMIR Publications

ABSTRACT

Background:

Tourette's Syndrome is a neurological disorder characterized by tics, which are repeated physical movement and vocal sounds. People with mild Tourette’s may experience a few tics throughout the day, while severe cases may have tics every five to ten seconds. Typically during high levels of stress, tics become chained in an incessant, continuous fashion—this is known as a tic attack. Tic attacks incapacitate the patient, rendering it difficult for them to move, speak, and perform daily actions. Caretakers can administer medication that can calm down tic attacks. This medication can be obtained faster if caretakers receive a notification when an attack occurs. Such a notification can be sent via TSBAND, a wearable wristband that we have developed which utilizes machine learning algorithms and a variety of sensors to detect tic attacks.

Objective:

The objectives of this study are to (1) explain the mechanisms of monitoring and detecting tic attacks; (2) present a physical device to monitor tic attacks and send notifications to a mobile application; (3) describe the need for a patient-to-caretaker bridge and exemplify possible techniques of accomplishing this through both automatic and manual communication mechanisms.

Methods:

In this study, a prototype of TSBAND was fully developed along with a companion mobile application, and an online survey was filled out by 70 Tourette’s patients to assess whether such a wristband would be helpful. TSBAND uses a triaxial accelerometer, gyroscope, heart rate sensor, and a body temperature and humidity sensor to detect attacks. Using machine learning models, the watch tailors the detection algorithm to the patient's daily routines and tic patterns. Vocal attacks can also be monitored using an algorithm that analyzes repetitions in speech. Notifications are sent to caretakers through a mobile application. There are also two backup mechanisms, a vocal cue phrase and a physical button, to manually notify caretakers. Notification requests can also be cancelled through the physical button.

Results:

In the survey, 91.4% of the patients said they struggled to communicate with caretakers during a tic attack, and 75.7% of the patients believed the device would be helpful. The audio algorithm led to a 92% accuracy rate, while the physical component has not been tested with patients yet. The prototype is an affordable solution, costing $62.74. We are now aiming to test the device with Tourette’s patients, improve the detection mechanism, and upgrade to higher-quality sensors by September 2023.

Conclusions:

This study presents an affordable and effective solution to automatically detect tic attacks in Tourette’s patients, giving them faster access to support and medication. As the detection mechanism is customized, TSBAND can help not only Tourette’s patients, but can potentially be expanded to support a variety of tremor patients.