Artificial intelligence methods for identification of ADHD in children based on EEG signals

Abstract

Children with Attention Deficit Hyperactivity Disorder (ADHD) are among the most common neurodevelopmental disorders. The incidence of this disorder in society shows an increasing trend from the past to the present. Recent developments suggest that Artificial Intelligence and Electroencephalogram (EEG) analysis can accurately diagnose cases of ADHD in children. By combining a new type of Continuous Wavelet Transform (CWT) with Variational Mode Decomposition (VMD), a novel algorithm for self-adaptive signal processing that is more resilient to sampling and noise, the suggested method decomposed EEG signals using detection and removal of noise and extraction of relevant features. The classifier that uses Convolutional Neural Networks (CNNs) and Bidirectional Long Short-Term Memory (BiLSTM) networks, the study's Deep Learning (DL) algorithm used the EEG waves as input data. Results: An algorithm has been proposed that distinguishes between approximately 94% of individuals using a 17-channel EEG signal to compare healthy individuals with those who have ADHD. The proposed method, using the CNN-BiLSTM method to analyze EEG signals and process the data in a DL algorithm produced a classification accuracy of 98.69%. The combination of precise EEG with AI methods holds promise for improving our understanding of ADHD in children and developing more accurate diagnostic tools.