Recurrent and Convolutional Neural Networks in Classification of EEG Signal for Guided Imagery and Mental Workload Detection

TL;DR

The paper addresses binary EEG classification of Guided Imagery relaxation versus mental workload and compares four deep learning architectures (EEGNet, LSTM, 1D-CNN, 1D-CNN-LSTM) on two data configurations (FULL-256 and COGN-26) using 1-second segments with minimal preprocessing. EEGNet serves as a reference, while 1D-CNN and the hybrid 1D-CNN-LSTM frequently outperform it, especially on the cognitive electrode subset; LSTM alone performs worst among the tested models. On full 256-channel data, 1D-CNN-LSTM achieves the highest cross-validated accuracy (0.7726), with 1D-CNN close behind (0.7682), and EEGNet at 0.7615; on the 26-electrode set, 1D-CNN reaches 0.8094 accuracy, outperforming the others. The results show that restricting to a targeted 26-electrode set can match or exceed full-capacitance performance, suggesting that simpler setups with raw EEG can achieve strong discrimination without manual feature extraction, which has practical implications for wearable EEG-based relaxation and workload monitoring systems.

Abstract

The Guided Imagery technique is reported to be used by therapists all over the world in order to increase the comfort of patients suffering from a variety of disorders from mental to oncology ones and proved to be successful in numerous of ways. Possible support for the therapists can be estimation of the time at which subject goes into deep relaxation. This paper presents the results of the investigations of a cohort of 26 students exposed to Guided Imagery relaxation technique and mental task workloads conducted with the use of dense array electroencephalographic amplifier. The research reported herein aimed at verification whether it is possible to detect differences between those two states and to classify them using deep learning methods and recurrent neural networks such as EEGNet, Long Short-Term Memory-based classifier, 1D Convolutional Neural Network and hybrid model of 1D Convolutional Neural Network and Long Short-Term Memory. The data processing pipeline was presented from the data acquisition, through the initial data cleaning, preprocessing and postprocessing. The classification was based on two datasets: one of them using 26 so-called cognitive electrodes and the other one using signal collected from 256 channels. So far there have not been such comparisons in the application being discussed. The classification results are presented by the validation metrics such as: accuracy, recall, precision, F1-score and loss for each case. It turned out that it is not necessary to collect signals from all electrodes as classification of the cognitive ones gives the results similar to those obtained for the full signal and extending input to 256 channels does not add much value. In Disscussion there were proposed an optimal classifier as well as some suggestions concerning the prospective development of the project.

Figures (9)

• Figure 1: On the left: EGI 256-channel EEG cap.On the right: the overview of the whole EEG Laboratory at UMCS, Lublin, Poland
• Figure 2: Electrodes placement on HydroCel GSN 130 Geodesic Sensor Netgeodesics2009geodesicwojcik2023investigating
• Figure 3: Power spectral density of different frequency bands shown for 1-s segment of signal from GI sample subject
• Figure 4: Power spectral density of different frequency bands shown for 1-s segment of signal from MT sample subject
• Figure 5: Data science pipeline - steps of data preparation for training