Analysis of EEG brain connectivity of children with ADHD using graph theory and directional information transfer

Ali Ekhlasi,Ali Motie Nasrabadi,Mohammadreza Mohammadi
DOI: https://doi.org/10.1515/bmt-2022-0100
2022-10-07
Biomedical Engineering / Biomedizinische Technik
Abstract:Research shows that Attention Deficit Hyperactivity Disorder (ADHD) is related to a disorder in brain networks. The purpose of this study is to use an effective connectivity measure and graph theory to examine the impairments of brain connectivity in ADHD. Weighted directed graphs based on electroencephalography (EEG) signals of 61 children with ADHD and 60 healthy children were constructed. The edges between two nodes (electrodes) were calculated by Phase Transfer Entropy (PTE). PTE is calculated for five frequency bands: delta, theta, alpha, beta, and gamma. The graph theory measures were divided into two categories: global and local. Statistical analysis with global measures indicates that in children with ADHD, the segregation of brain connectivity increases while the integration of the brain connectivity decreases compared to healthy children. These brain network differences were identified in the delta and theta frequency bands. The classification accuracy of 89.4% is obtained for both in-degree and strength measures in the theta band. Our result indicated local graph measures classified ADHD and healthy subjects with accuracy of 91.2 and 90% in theta and delta bands, respectively. Our analysis may provide a new understanding of the differences in the EEG brain network of children with ADHD and healthy children.
What problem does this paper attempt to address?