Real-Time Epileptic Seizure Detection During Sleep Using Passive Infrared (Pir) Sensors

According to World Health Organization (WHO), millions of people suffer from epilepsy, which is a chronic disorder of the brain. Sudden Unexplained Death in Epilepsy (SUDEP) is considered as one of the most dangerous threats to the patients who suffer from convulsive seizures during sleep. In this presentation we will demonstrate how we can detect epileptic seizures during sleep using Passive Infrared Sensors (PIR). The causes of the SUDEP are still unknown, and the patients with epileptic seizures can experience death in bed, likely after an unwitnessed convulsive seizure. Alerting caregivers to intervene during a seizure will help to prevent the SUDEP. Most of the current approaches to detect epileptic seizures at home rely on contact-based sensors and cameras. We developed a privacy-safe, non-contact PIR sensor based system that can alert caregivers during a convulsive seizure [1]. During a convulsive seizure the whole body experiences rapid, rhythmic and sometimes violent shaking movements, often with loss of consciousness. The PIR sensor monitoring the sleeping person generates a chaotic looking signal during a seizure. This signal can be distinguished from other ordinary body movements using a machine learning algorithm. We developed an algorithm using Hidden Markov models (HMM) and a convolutional neural network to classify the sensor data set into either occurrence or non-occurrence of epileptic convulsive seizures. The PIR sensor-based seizure detection system is a contact-free system. Epileptic seizures can be detected from 1 meter up to 11 meters distance between the sensor and the subject. During the demonstration, we will simulate a seizure by shaking our body in front of our system and generate an alarm. Reference: [1] Hanosh, O., Ansari, R., Younis, Cetin, AE (2019). Real-Time Epileptic Seizure Detection during Sleep using Passive Infrared Sensors. IEEE Sensors Journal.