Comprehensive review of EEG data classification techniques for ADHD detection using machine learning and deep learning

Abstract

Children who have Attention-Deficit/Hyperactivity Disorder (ADHD) have a chronic behavioral disease. Children with ADHD have a hard time focusing and controlling their actions. One of the most difficult problems in controlling and treating this condition is early detection. There is yet to be discovered a reliable professional procedure for early detection of this condition. The electroencephalogram (EEG) is a useful neuroimaging technique for researching ADHD; one of the key goals is to define the EEG of ADHD youngsters. Numerous methods based on EEG signals have been put out in the literature to address this issue since they are an effective neuroimaging approach for studying ADHD. The best recording formats and channels for diagnosing ADHD, however, have not been the subject of many research. Machine learning (ML) and Artificial Intelligence (AI) strategies for identifying ADHD using EEG-based tools are discussed in this paper. Although, in the case of ADHD, the utilization of ML and AI approaches is restricted. However, the data clearly imply that combining EEG technologies with ML/AI may be utilized to detect ADHD. For categorizing adult ADHD subtypes based on EEG power spectra, ML algorithms that incorporate several classifiers are presented. A widely used deep learning (DL) method is the convolutional neural network (CNN). The use of DL approaches in ADHD research, on the other hand, is currently restricted. EEG has been used in studies to look for ADHD neurological connections. Recent advances in deep learning algorithms, particularly CNN, are anticipated to overcome the issue.