A Fast and Accurate Approach for Real-Time Seizure Detection in the IoMT

Abstract

We propose an EEG-based seizure detection method which uses the discrete wavelet transform (DWT), Hjorth parameters and a k-NN classifier. Seizure detection is performed in three stages. In the first stage, EEG signals are decomposed by the DWT into sub-bands and Hjorth parameters are extracted from each of these sub-bands. In the second stage, a k-NN classifier is used to classify the EEG data. The results demonstrate a significant difference in Hjorth parameters between interictal and ictal EEG with ictal EEG being less complex than interictal EEG. We report an accuracy of 100% for a classification of normal vs. ictal EEG and 97.9% for normal and interictal vs. ictal EEG. We propose an Internet of Medical Things (IoMT) platform for performing seizure detection. The proposed framework accommodates the proposed scheme for seizure detection and allows communication of detection results. The IoMT framework also allows the adjustment of seizure detection parameters in response to updated performance evaluations, and possible changes in seizure and signal characteristics as well as the incorporation of other sensor signals to provide an adaptive, multi-modal framework for detecting seizures.