Detection of Attention Deficit /Hyperactivity Disorder in Adults Using Complexity-Based Electroencephalography Features

Attention Deficit Hyperactivity Disorder (ADHD) is a neurodevelopmental condition that is characterized by inattention, over-activity, and impulsivity. This paper proposed a framework for evaluating complexity variations and classifying ADHD adults from control participants using non-linear dynamics in Electroencephalogram (EEG) signals during rest. For this aim, 19-channel EEG signal was recorded from 30 adult ADHD and 30 control participants. After pre-processing recorded signals, Lempel-Ziv Complexity (LZC), relative power, and power ratios were extracted from them in delta, theta, alpha, beta, and gamma frequency bands. ANOVA statistical test has been used to investigate differences in the extracted features between the two groups. Also, the Support Vector Machine (SVM) was used to evaluate the ability to classify statistically significant features. The results indicated that the EEG complexity of the adult ADHD group is higher than neurotypical controls, especially in the temporal electrodes with a p-value $\lt 0.01$. Also, combining complexitybased and frequency-based features has the best classification performance with $78.02 \pm 1.80 \%$ accuracy.