Quantitative and Qualitative Comparison of EEG-Based Neural Network Organization in Two Schizophrenia Groups Differing in the Duration of Illness and Disease Burden: Graph Analysis With Application of the Minimum Spanning Tree

• Published in: Clinical EEG and neuroscience 2019-07, Vol.50 (4), p.231-241
• Main Authors: Jonak, Kamil, Krukow, Paweł, Jonak, Katarzyna E., Grochowski, Cezary, Karakuła-Juchnowicz, Hanna
• Format: Article
• Language: English
• Subjects: Adult; Brain - physiopathology; EEG; Electroencephalography; Female; Graph theory; Humans; Illnesses; Image Processing, Computer-Assisted; Male; Mental disorders; Models, Neurological; Network topologies; Neural networks; Neural Pathways - physiopathology; Qualitative analysis; Qualitative research; Schizophrenia; Schizophrenia - physiopathology; Signal Processing, Computer-Assisted; Topology; Trees
• ISSN: 1550-0594; 2169-5202
• DOI: 10.1177/1550059418807372

The aim of this study was to compare neural network topology of 30 patients with first episode schizophrenia (FES) and 30 multiepisode schizophrenia (mean number of psychotic relapses =4 years, duration of illness >5 years) patients, who were assessed with graph theory methods. This comparison was designed to identify network differences, which might be assigned to the burden of a mental disease. To estimate functional connectivity, we applied the phase lag index algorithm and the minimum spanning tree (MST) for the characterization of network topology. Group comparison revealed significant between-group differences of maximal betweenness centrality and tree hierarchy in the beta-band and hierarchy in the gamma-band. MST results showed that in the beta-band the network of patients with longer duration of illness (LDI) was characterized by more centralized network, while subjects with short duration of illness (FES) showed more decentralized topology. Furthermore, in the gamma-band, our results suggest that illness duration can disturb the balance between overload prevention and large-scale integration in the brain network. A qualitative analysis proved that the topological displacement of hubs also differentiated the FES and LDI groups. Our findings suggest that the duration of illness significantly affects the topology of resting-state functional network, supporting the “disconnectivity hypothesis’ in schizophrenia.