Applicable features of electroencephalogram for ADHD diagnosis

Abstract

Purpose Attention-deficit/hyperactivity disorder (ADHD) is a neuro-developmental and psychiatric disorder, which affects 11% of children around the world. Several linear and nonlinear biomarkers from electroencephalogram (EEG) signals have been proposed for diagnosis of ADHD to date. However, the determination of which type of analysis gives us the best feature and biomarker to diagnose ADHD is still controversial. In this study, we aimed to evaluate and compare several categories of features, extracted from EEG signals, for diagnosis of ADHD. Methods Thirty 7-12-year-old children fulfilling the DSM5 criteria for ADHD and thirty healthy children underwent a noninvasive EEG evaluation at resting-state. After preprocessing, five categories of features including morphological, time, frequency, time-frequency, and nonlinear features were extracted from EEGs. The efficacy of each feature category in ADHD diagnosis was determined using statistical analysis, receiver operating characteristic (ROC) curves, and evidential K-nearest neighbor (EKNN) classifier. Results Statistical analyses showed that 13.15, 13.68, 14.47, 14.03, and 34.73% of extracted features were significant (p < 0.05) in morphological, time, frequency, time-frequency, and nonlinear domains, respectively. The largest AUC values for the five morphological, time, frequency, time-frequency, and nonlinear feature categories, were 0.870, 0.796, 0.824, 0.806, and 0.899, respectively. We obtained the accuracies of 77.43% using morphological features, 74.09% using time features, 80.44% using frequency features, 78.50% using time-frequency features, and 86.40% using nonlinear features. Conclusion Our results showed that EEG nonlinear analysis is a good quantitative tool to detect the abnormalities of the electrical activity of the brain in ADHD. This result was expected due to the complexity of the brain and the nonlinear nature of the EEG signal. Therefore, better outcomes may be expected in early diagnosis of diseases, especially psychiatric disorders, by increasing the use of nonlinear methods.

EEG Analysis with Wavelet Transform under Music Perception Stimulation

“…Statistical analysis showed that 13.15, 13.68, 14.47, 14.03, and 34.73% of the extracted features were significant in morphology, time, frequency, time frequency, and nonlinear domain ( P < 0.05). The maximum AUC values of five categories such as morphological, temporal, frequency, time frequency, and nonlinear feature are 0.870, 0.796, 0.824, 0.806, and 0.899 [ 8 ]. Wavelet packet decomposition reconstruction of affective EEG signals and the ß rhythm was used for affective state recognition.…”

Computational Neuroscience Approach to Psychiatry: A Review on Theory-driven Approaches

“…This can be supported by a more focus on establishing utility through the active application of computational appro-aches in clinical trials. Combining theory- and data-driven approaches can be an appropriate way [ 14 , 75 - 79 ]. Combining theory- and data-driven approaches can be very helpful from an applied point of view [ 80 ].…”

Windowing Compensation in Fourier Based Surrogate Analysis and Application to EEG Signal Classification

“…Larger tests should be performed, not only to increase the confidence in method and allow meta parameter optimization, but also to test the performances of other nonlinear features in the MSWSA method. The selection of the nonlinear feature should include those commonly studied for ADHD [56][57][58], ADD [59] and ANX [60]. Also, only the MSWSA was used for classification to show that the featured contained information useful for classification purpose.…”