A Novel Hybrid Deep Neural Network Classifier for EEG Emotional Brain Signals

The field of brain computer interface (BCI) is one of the most exciting areas in the field of scientific research, as it can overlap with all fields that need intelligent control, especially the field of the medical industry. In order to deal with the brain and its different signals, there are many ways to collect a dataset of brain signals, the most important of which is the collection of signals using the non-invasive EEG method. This group of data that has been collected must be classified, and the features affecting changes in it must be selected to become useful for use in different control capabilities. Due to the need for some fields used in BCI to have high accuracy and speed in order to comply with the environment's motion sequences, this paper explores the classification of brain signals for their usage as control signals in Brain Computer Interface research, with the aim of integrating them into different control systems. The objective of the study is to investigate the EEG brain signal classification using different techniques such as Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), as well as the machine learning approach represented by the Support Vector Machine (SVM). We also present a novel hybrid classification technique called CNN-LSTM which combines CNNs with LSTM networks. This proposed model processes the input data through one or more of the CNN's convolutional layers to identify spatial patterns and the output is fed into the LSTM layers to capture temporal dependencies and sequential patterns. This proposed combination modelling to achieve high efficacy across domains. A test was done to determine the most effective approach for classifying emotional brain signals that indicate the user's emotional state. The dataset used in this research was generated from a widely available MUSE EEG headgear with four dry extra-cranial electrodes. The comparison came in favor of the proposed hybrid model (CNNLSTM) in first place with an accuracy of 98.5% and a step speed of 244 milliseconds/step; the CNN model came in the second place with an accuracy of 98.03% and a step speed of 58 milliseconds/step; and in the third place, the LSTM model recorded an accuracy of 97.35% and a step speed of 2 sec/step; milliseconds/step running speed.