Abstract
We design an algorithm to automatically detect epileptic seizure onsets and offsets from scalp electroencephalograms (EEGs). The proposed scheme consists of two sequential steps: detecting seizure episodes from long EEG recordings, and determining seizure onsets and offsets of the detected episodes. We introduce a neural network-based model called ScoreNet to carry out the second step by better predicting the seizure probability of pre-detected seizure epochs to determine seizure onsets and offsets. A cost function called log-dice loss with a similar meaning to the F1 score is proposed to handle the natural data imbalance inherent in EEG signals signifying seizure events. ScoreNet is then verified on the CHB-MIT Scalp EEG database in combination with several classifiers including random forest, convolutional neural network (CNN), and logistic regression. As a result, ScoreNet improves seizure detection performance over lone epoch-based seizure classification methods; F1 scores increase significantly from 16-37% to 53-70%, and false positive rates per hour decrease from 0.53-5.24 to 0.05-0.61. This method provides clinically acceptable latencies of detecting seizure onset and offset of less than 10 seconds. In addition, an effective latency index is proposed as a metric for detection latency whose scoring considers undetected events to provide better insight into onset and offset detection than conventional time-based metrics.
Highlights
A N epileptic seizure can be defined as a transient event of abnormal electrical activity in the brain [1]
To remove human errors from EEG-based seizure detection, we develop an automated epileptic seizure detection system that can label the time of seizure events in EEG signals
We address this issue by establishing a cost function in ScoreNet called log-dice loss based on a dice similarity coefficient
Summary
A N epileptic seizure can be defined as a transient event of abnormal electrical activity in the brain [1]. A CNN was proposed to spatially and temporally capture ictal patterns in EEG epochs [34] This method included an additional post-processing procedure designed from clinical characteristics of seizures to reduce the occurrence of false positives and false negatives from prior classification. Since EEG seizure data are naturally highly imbalanced, the outcomes of existing detection methods tend to be biased towards a normal class We address this issue by establishing a cost function in ScoreNet called log-dice loss based on a dice similarity coefficient. Another issue in the abovementioned studies [33], [35] was that the mean latencies used for interpreting seizure onsetoffsets were misleading; positive and negative latencies were defined as early and late predictions of seizure onset-offsets, respectively, and they could cancel each other out during the calculation of mean latencies. This article is organized as follows: Section II presents the process of seizure onset-offset detection; Section II-B provides an in-depth explanation of ScoreNet; Section II-C provides the problem formulation including the proposed loss function; Section III describes the EL-index; Section IV outlines all the experiments conducted to verify ScoreNet; Section V presents the seizure classification and seizure onset-offset determination results, accompanied by discussions and with graphical illustrations
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: IEEE Transactions on Neural Systems and Rehabilitation Engineering
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
