Development of Clinical Detection System for Tonic-Clonic Seizures using New Wearable Sensors

Abstract

This paper presents the development and testing of a new wireless body area network (WBAN) designed specifically for remote patient monitoring for detection of generalized tonic-clonic seizures (GTCS). The WBAN is comprised of four, triaxial accelerometer sensors worn on arms and legs. Commercial medical sensor modules were used as benchmarks to validate and calibrate the wearable sensors data. Each sensor is communicating with Linux based platform via Bluetooth Low Energy (BLE) network. Graphical User Interface (GUI) was developed using Python Software to display the collected real-time sensors data. A new digital signal processing routine for seizure detection was created using Matlab software. This routine is employing multi-feature techniques based on Fast Fourier Transform (FFT) and moving average window to analyze data samples with minimal processing time. The average windowed samples are compared periodically with pre-set empirical threshold values every second on each axis for the four sensor nodes. An alert is generated once seizure is detected at base station where Short Message Service (SMS) and instant email will be sent to the medical health staff. The real-time data is stored in secured medical database for future reference and analysis. The final WBAN prototype was demonstrated using healthy volunteers mimicking tonic-clonic seizure symptoms to test overall system performance. The system is also approved for future clinical trials in Canada, and pilot clinical demonstrations have been achieved.