Master thesis : Creation of test procedure for the analysis of signal quality in the comparison of wearables against gold standard medical devices for the prevention of epilepsy

Abstract

Nowadays, the use of wearable medical devices is becoming more and more common and is a popular solution for patients with medical conditions. These can be used to monitor and control neurological disorder as Parkinson’s, Alzheimer’s, migraine, epilepsy... This project focuses on the use of wearables in epilepsy which affects about 50 million people worldwide. More particularly, it is about the efficiency of the wearable device to detect Generalized tonic–clonic seizure (GTCS). It is now possible to detect seizures through the analysis and processing of different biosignals recorded in the wearable thanks to the accelerometer, the photoplethysmograph, and the electrodermal activity. Thanks to the combination of them, it is possible to detect seizures and prevent patients from dangerous seizure consequences. The main objective of this thesis project is to create a test procedure for the analysis of signal quality in the comparison of wearable versus gold standard medical devices with the aim and purpose of predicting epileptic seizures with the most accurate wearable device avail- able on the market. The creation of the experimental test protocol aims to assess the signal quality of sensors and the accuracy of wearable devices such as smartwatches in detecting GTCS by comparing them to the gold standard that has already proven its effectiveness in the medical context. In this work, extensive research on the background of the disease and on the biosignals related has been done. Then, there is an overview of the state of the art in the available devices to monitor epilepsy, whether invasive or non-invasive. All this information is gathered to create the final protocol separated into two main parts: the signal quality assessment (performed with four rules) and the seizure detection accuracy assessment (comparison to state of the art detection system). The protocol is mainly based on the literature in recent years. Finally, this procedure is summarized in the shape of a Standard Operating Procedure that can be used in the future by a third party. This project is not a finished project but is intended to serve as a solid foundation for a larger project.