A Deep Learning-Based Wearable Bracelet for Real-Time Clonic Seizure Detection Using Accelerometer Data

About 50 million people worldwide suffer from epilepsy, a neurological disorder characterized by spontaneous seizures resulting from abnormal brain activity. Existing seizure detection devices often face limitations such as discomfort, unreliability, or inaccuracy, leading to issues such as high cost or inadequate care. In response, this study presents an affordable and accurate wearable bracelet for clonic epileptic seizure detection using deep learning. The bracelet integrates an accelerometer sensor and a Bluetooth module to record patient movements. The recorded data are transmitted via Bluetooth to a MATLAB application, where a Convolutional Neural Network (CNN) classifier distinguishes seizure-like activity from normal movements with an accuracy of 94.12 %. Upon detecting a seizure, the system immediately notifies designated caregivers to support timely intervention. Compared to existing detection methods, the proposed wearable system offers improved accuracy and real-time caregiver notification. The proposed system provides immediate support and has the potential to improve seizure management and overall quality of life for individuals with epilepsy.