Abstract
Epilepsy is a long-term neurogenic disease that requires caregivers to accompany the patient days and nights. Caregivers have to help the patients immediately when they are having a seizure, which could cause vital injuries or even death. To address this issue, we designed a bracelet containing a three-dimensional accelerometer and a three-dimensional gyroscope to record the movements of the patient and built a Random Forest model to automatically detect seizures in at most 10 seconds upfront. We designed a home-based data-collecting method that allows patients to stay at home or perform their daily activities outside the hospital. Data collected in this method would be similar to the situation in which the patients would actually use wearable monitoring devices at home. The performance was evaluated based on an experimental study of epilepsy detection and classification, where epileptic motor data was collected in the West China Hospital of Sichuan University. Due to the experimental results, our daytime seizure detection model achieved 75.91% sensitivity and 88.90% precision, while our nighttime seizure detection model achieved 88.01% sensitivity and 88.33% precision. These preliminary results indicate that this home-based data collection method can capture seizures efficiently.
Original language | English |
---|---|
Title of host publication | 2021 10th International IEEE/EMBS Conference on Neural Engineering, NER 2021 |
Publisher | Institute of Electrical and Electronics Engineers |
Pages | 854-857 |
Number of pages | 4 |
ISBN (Electronic) | 9781728143378 |
DOIs | |
Publication status | Published - May 2021 |
Event | 10th International IEEE EMBS Conference on Neural Engineering (VIRTUAL) - Duration: 4 May 2021 → 6 May 2021 |
Conference
Conference | 10th International IEEE EMBS Conference on Neural Engineering (VIRTUAL) |
---|---|
Period | 4/05/21 → 6/05/21 |