There are various depressive subtypes identified in patients with major depressive disorder (MDD). Depression with psychotic symptoms is usually known to be a severe type of depression that includes symptoms such as delusions and/or hallucinations, and remains a common condition that is often underrecognized and inadequately treated in clinical practice. Electroencephalography (EEG) biomarkers have been implicated to classify healthy and psychopathological neural signals using machine learning algorithms. In this study, we sought to identify cortical functional connectivity metrics that differentiate network manifestation of different depressive subtypes and healthy controls. We first performed replication analyses to obtain the principal functional connectivity microstates across each independent group (healthy controls, psychotic depressions and nonpsychotic depressions). Next, we examined temporal functional connectivity dynamics in each group. The results show that fundamental dynamic functional connectivity microstates are highly reproducible, both within and across participants. Based on the temporal and sequential parameters (mean duration, fractional windows and transition number) derived from dynamic functional connectivity analysis, we found inter-group differences across healthy and MDD subgroups statistically significant. These results show that the principal FC microstates dynamics are essential neural biomarkers distinctly associated with depression clinical phenotypes.Clinical relevance-Our findings suggest that a network-level feature, that may reflect the neurobiological difference between different depression subtypes, and healthy controls, and in turn may contribute towards a scalable EEG-based assisted diagnostic tool.
Read full abstract