Classification of 1D and 2D EEG Signals for Seizure Detection in the Newborn Using Convolutional Neural Networks

Abstract

Unlike adults, neonates do not always show clinical symptoms during seizures.\rTherefore, uncontrolled seizures lead to severe brain damage. Timely recognition of\rseizures plays a crucial role for neonates. In this study, a deep transfer learning\rapproach was proposed for automatic detection of seizures on the C4-P4 channel\rusing electroencephalography (EEG) signals from neonates. The EEG signals were\rused in 1D and 2D dimensions to ensure performance, robust functionality, and a\rclinically acceptable level of detection accuracy. The pre-trained deep learning\rmodels Alexnet, ResNet, GoogleNet and VggNet were used in the study.\rSpectrograms were obtained by converting 1-dimensional signal data into 2-\rdimensional images, and then classification was performed for both the 1D and 2D\rdatasets. For 1D classification, the highest performance was obtained by VggNet\rarchitecture with 91.67%, while 2D classification was obtained by AlexNet and\rResNet architecture with 95.83%. The use of spectrograms significantly improved\rclassification performance and made neonatal seizure detection and decision-making\rmore clinically reliable.