Deep Convolutional Neural Network Regularization for Alcoholism Detection Using EEG Signals

Abstract

Alcoholism is attributed to regular or excessive drinking of alcohol and leads to the disturbance of the neuronal system in the human brain. This results in certain malfunctioning of neurons that can be detected by an electroencephalogram (EEG) using several electrodes on a human skull at appropriate positions. It is of great interest to be able to classify an EEG activity as that of a normal person or an alcoholic person using data from the minimum possible electrodes (or channels). Due to the complex nature of EEG signals, accurate classification of alcoholism using only a small dataset is a challenging task. Artificial neural networks, specifically convolutional neural networks (CNNs), provide efficient and accurate results in various pattern-based classification problems. In this work, we apply CNN on raw EEG data and demonstrate how we achieved 98% average accuracy by optimizing a baseline CNN model and outperforming its results in a range of performance evaluation metrics on the University of California at Irvine Machine Learning (UCI-ML) EEG dataset. This article explains the stepwise improvement of the baseline model using the dropout, batch normalization, and kernel regularization techniques and provides a comparison of the two models that can be beneficial for aspiring practitioners who aim to develop similar classification models in CNN. A performance comparison is also provided with other approaches using the same dataset.

An Approach for Pronunciation Classification of Classical Arabic Phonemes Using Deep Learning

“…This approach helps construct CNN models from scratch and improves them by applying various techniques of data augmentation, finetuning neural networks, and hyperparameters tuning. Similar methods are used in other domains [34] for CNN models improvements. Given that the model's performance improved with synthetic data, there are chances of achieving high accuracy if more real data can be retrieved.…”

“…With our constructed dataset and the first CNN model, we achieved a training accuracy of 84.27% and validation accuracy of 40.15%. Based on previous approaches [34,35], and the low validation accuracy, it was evident that the small dataset could not give high classification accuracy. Because obtaining more user data was not an option, for achieving good accuracy, we developed a two-phase approach as shown in Figure 1.…”

Deep Convolutional Neural Network-Based Visual Stimuli Classification Using Electroencephalography Signals of Healthy and Alzheimer’s Disease Subjects

“…After each convolution, batch normalization is performed to achieve model stability. Additionally, dropout layers are used to significantly reduce overfitting [ 26 ]. The final multi-class classification layer uses the SoftMax function [ 57 ].…”

“…To improve the best-performing model EEGNet SSVEP, the L1 and L2 regularization methods were added to the final fully connected layer with both regularization penalty values equal to 0.001. The addition of constraints to the model weights has been shown to minimize the complexity of the model [ 26 ]. One study improved the accuracy of identifying motor movements from the EEG data by 2% using the EEGNet model with regularization [ 56 ].…”

“…In recent years, deep neural networks (DNNs) have received increasing attention from researchers for a variety of classification tasks by using EEG data: alcoholism detection [ 26 ], predicting early stages of schizophrenia [ 27 ], classifying motor imagery to assist brain–computer interfaces [ 6 , 28 ], determining the stage of AD [ 29 ], and even the stages of visual processing [ 30 ]. The growing interest in visual perception may open up more opportunities to adapt BCI systems to visually impaired people [ 30 ].…”