The Reliability of Diagnosing Schizophrenia Using the GRU Layer in Conjunction with EEG Rhythms

Pankaj Kumar  Sahu; Karan  Jain

doi:10.18502/fbt.v12i3.19184

Pankaj Kumar Sahu Department of Instrumentation and Control Engineering, Dr. BR Ambedkar National Institute of Technology Jalandhar, Punjab, India
Karan Jain Department of Instrumentation and Control Engineering, Dr. BR Ambedkar National Institute of Technology Jalandhar, Punjab, India

DOI: https://doi.org/10.18502/fbt.v12i3.19184

Keywords: Gated Recurrent Unit; Rudiment Densely-Coupled Convolutional; Densely-Coupled Convolutional; 1-D-Convolution Layer.

Abstract

Purpose: At resting state, the human brain releases cycles of Electroencephalography (EEG), which has been proven aberrant in persons with schizophrenia. Deep learning methods and patterns found in EEG of brain activity are helpful features for verifying schizophrenia. The proposed study demonstrates the applicability of alpha-EEG rhythm in a Gated-Recurrent-Unit-based deep learning model for studying schizophrenia.

Materials and Methods: This study suggests Rudiment Densely-Coupled Convolutional Gated Recurrent Unit (RDCGRU) for the EEG rhythm (gamma, beta, alpha, theta, and delta) based diagnoses of schizophrenia. The model includes multiple 1-D-Convolution (Con-1-D) folds with steps greater than 1, which enables the model to programmatically and effectively learn how to reduce the incoming signal. The Con-1-D layers and numerous Gated Recurrent Unit (GRU) layers comprise the Exponential-Linear-Unit activation function. This powerful activation function facilitates in-deep-network training and improves classification performance. The Densely-Coupled Convolutional Gated Recurrent Unit (DCGRU) layers enable RDCGRU to address the training accuracy loss brought on by vanishing or exploding gradients. This makes it possible to develop intense, deep versions of RDCGRU for more complex problems. The sigmoid activation function is implemented in the digital classifier's output nodes.

Results: The RDCGRU framework performs efficiently with alpha-EEG rhythm (88.06%) and harshly with delta-EEG rhythm (60.05%). The research achievements: In EEG rhythm-based schizophrenia verification, GRU cells with the RDCGRU deep learning model performed better with alpha-EEG rhythm.

Conclusion: The α-EEG rhythm is a crucial component of the RDCGRU deep learning model for studying schizophrenia using EEG rhythms. In our investigation of RDCGRU deep learning architectures, we noticed that Con-1-D layers connected special learning networks function well with the α-EEG rhythm for the EEG rhythm-based verification of schizophrenia.