Application of a Low-Cost mHealth Solution for the Remote Monitoring of Patients With Epilepsy: Algorithm Development...

Area of Science:

• Neurology
• Biomedical Engineering
• Mobile Health

Background:

• Automated seizure detection in long-term electroencephalography (EEG) analysis is crucial for remote patient monitoring in epilepsy.
• Current methods may lack accessibility and real-time capabilities for widespread patient use.

Purpose of the Study:

• To investigate the feasibility of using smartphones for processing large EEG recordings for remote epilepsy patient monitoring.
• To develop and evaluate a mobile health (mHealth) solution for automated seizure detection.

Main Methods:

• Developed the Sezect Android application using the Chaquopy SDK for automated EEG analysis and seizure classification.
• Employed a cross-database model with specific features (successive decomposition index, matrix determinant) and adaptive baseline correction.
• Utilized a postprocessing-based support vector machine classifier across five EEG databases and tested on various smartphone models.

Main Results:

• Achieved high performance with 93.5% sensitivity and 97.5% specificity.
• Demonstrated a low false detection rate of 1.5 per hour.
• Found minimal variation in processing time across different smartphone models, indicating broad applicability.

Conclusions:

• The Sezect mHealth app is a valuable tool for real-time EEG data processing and seizure detection in epilepsy management.
• Smartphone-based EEG analysis presents a viable and accessible approach for remote patient monitoring, enhancing patient care and quality of life.