Detection of epileptic seizure using wireless sensor networks

Area of Science:

• Biomedical Engineering
• Neurology
• Wearable Technology

Background:

• Current electroencephalogram (EEG) monitoring for epilepsy is accurate but uncomfortable, restricting patient movement and long-term home use.
• The need for non-invasive, comfortable, and mobile seizure monitoring solutions is critical for improving patient quality of life.
• Wireless sensor networks offer a promising approach for unobtrusive, continuous patient monitoring.

Purpose of the Study:

• To develop and evaluate a novel seizure detection system utilizing wireless accelerometry.
• To assess the efficacy of the K-Nearest Neighbor (KNN) and Artificial Neural Network (ANN) algorithms for classifying seizure movements.
• To determine the feasibility of a system for remote patient monitoring in clinical or home settings.

Main Methods:

• Deployment of three 2D wireless accelerometer sensors on the right arm, left arm, and left thigh of epilepsy patients.
• Development of an automatic detection algorithm using data from three severe epilepsy patients.
• Analysis of sensor data using K-Nearest Neighbor (KNN) and Artificial Neural Network (ANN) for movement classification.

Main Results:

• The K-Nearest Neighbor algorithm demonstrated superior performance compared to the Artificial Neural Network for seizure detection.
• Accurate seizure detection was achieved when at least 50% of the signal comprised seizure samples.
• The system successfully identified seizure movements from normal daily activities, indicating high specificity.

Conclusions:

• Wireless accelerometry presents a viable and patient-friendly alternative to traditional EEG monitoring for epilepsy.
• The proposed system, utilizing KNN, offers accurate and automatic seizure detection without requiring patient-specific training.
• This technology has the potential to enhance patient independence and enable continuous monitoring in diverse environments.