Predicting epileptic seizures with a stacked long short-term memory network
Abstract
Despite advancements, seizure detection algorithms are trained using only the data recorded frompast epileptic seizures. This one-dimensional approach has led to an excessive false detection rate,where common movements are incorrectly classified. Therefore, a new method of detection isrequired that can distinguish between the movements observed during a generalized tonic-clonic(GTC) seizure and common everyday activities. For this study, eight healthy participants and twodiagnosed with epilepsy simulated a series of activities that share a similar set of spatialcoordinates with an epileptic seizure. We then trained a stacked, long short-term memory (LSTM)network to classify the different activities. Results show that our network successfullydifferentiated the types of movement, with an accuracy score of 94.45%. These findings present amore sophisticated method of detection that correlates a wearers movement against 12 seizurerelated activities prior to formulating a prediction.
Copyright (c) 2020 Jamie Pordoy, Ying Zhang, Nasser Matoorian, Massoud Zolgharni
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Copyright © by the authors; licensee Research Lake International Inc., Canada. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution Non-Commercial License (CC BY-NC) (http://creative-commons.org/licenses/by-nc/4.0/).