Alcohol use disorders automatic detection based BCI systems: a novel EEG classification based on machine learning and optimization algorithms

Abstract

Alcohol is a serious toxic substance that alters brain function by interfering with neuron processes in the central nervous system, leading to mental and behavioral disorders. Alcoholism has serious pathological effects on the liver, immune system, brain, and heart. The diagnosis of alcoholism is important not only because of its impact on individuals and society but also because of the cost to the national health system, as many people around the world suffer from the disease. These diseases can be diagnosed by automatically classifying normal and alcohol electroencephalogram (EEG) signals. Such that this work contains a simple and very fast prediction system. The method uses a bandpass filter to remove all unused signal frequencies, it can be showed in correlation matrices. Then the use of the machine learning algorithms in the classification stage that characterized by its high classification and prediction speed over than 400 and 25500 samples per second respectively, also, the use of optimization algorithms (GA and HHO) can be increased all accuracy values to more than 99.5% when using all electrodes and without importing data decomposition algorithms. To minimize the number of the electrodes and remained good accuracy values, this work uses the Extra-Trees algorithm, such as with only four electrodes the accuracy value remains higher of 99\%. A comparison with other techniques was performed aiming to validate our approach, and it shows great efficiency, simplicity, and instantly.