Pattern recognition (psychology)

Neuroscience

Clustering coefficient

Functional connectivity

Mathematics

Connectome

Resting state fMRI

Statistics

Betweenness centrality

Default mode network

Cognition

Artificial intelligence

Psychology

Functional magnetic resonance imaging

Connectomics

Major depressive disorder

Computer science

Centrality

Cluster analysis

Electroencephalography

IEEE Transactions on Neural Systems and Rehabilitation Engineering

Existing studies have shown functional brain networks in patients with major depressive disorder (MDD) have abnormal network topology structure. But the methods to construct brain network still exist some issues to be solved. This paper is to explore reliable and robust construction methods of functional brain network using different coupling methods and binarization approaches, based on high-density 128-channel resting state EEG recordings from 16 MDD patients and 16 normal controls (NC). It was found that the combination of imaginary part of coherence and cluster-span threshold outperformed other methods. Based on this combination, right hemisphere function deficiency, symmetry breaking and randomized network structure were found in MDD, which confirmed that MDD had aberrant cognitive processing. Furthermore, clustering coefficient in left central region in theta band and node betweenness centrality in right temporal region in alpha band were significantly negatively correlated with depressive level. And these network metrics had the ability to discriminate MDD from NC, which indicated that these network metrics might be served as the electrophysiological characteristics for probable MDD identification. Hence, this paper may provide reliable methods to construct functional brain network and offer potential biomarkers in MDD.

Graph Theory Analysis of Functional Connectivity in Major Depression Disorder With High-Density Resting State EEG Data