EEG Microstates in Mood and Anxiety Disorders: A Meta-analysis

Abstract

To reduce the psycho-social burden increasing attention has focused on brain abnormalities in the most prevalent and highly co-occurring neuropsychiatric disorders, such as mood and anxiety. However, high inter-study variability in these patients results in inconsistent and contradictory alterations in the fast temporal dynamics of large-scale networks as measured by EEG microstates. Thus, in this meta-analysis, we aim to investigate the consistency of these changes to better understand possible common neuro-dynamical mechanisms of these disorders.In the systematic search, twelve studies investigating EEG microstate changes in participants with mood and anxiety disorders and individuals with subclinical depression were included in this meta-analysis, adding up to 787 participants.The results suggest that EEG microstates consistently discriminate mood and anxiety impairments from the general population in patients and subclinical states. Specifically, we found a small significant effect size for B microstates in patients compared to healthy controls, with larger effect sizes for increased B presence in unmedicated patients with comorbidity. In a subgroup meta-analysis of ten mood disorder studies, microstate D showed a significant effect size for decreased presence. When investigating only the two anxiety disorder studies, we found a significantly small effect size for the increased microstate A and a medium effect size for decreased microstate E (one study). However, more studies are needed to elucidate whether these findings are diagnostic-specific markers.Results are discussed in relation to the functional meaning of microstates and possible contribution to an explanatory mechanism of overlapping symptomatology of mood and anxiety disorders.

Keywords: Anxiety; Bipolar Disorder; Depression; Meta-Analysis; Microstates; PTSD; Panic Disorder.

Fig. 1

PRISMA diagram describing the study selection process

Fig. 2

Forest plots of meta-analysis results for each microstate and temporal parameter. Left side - mean duration, right side - occurrence; Hedges’ g (squares proportional to weights used in meta-analysis) and associated confidence intervals (lateral tips) for individual data sets; summary measures (diamonds) of the whole group, anxiety subgroup that included two studies (Kikuchi et al. ; Terpou et al. 2022), and the mood subgroup that consists of ten studies (Bissonnette et al. ; Chen et al. ; Damborská et al. , ; He et al. ; Murphy et al. ; Qin et al. ; Sun et al. ; Wang et al. ; Zhao et al. 2022)

Fig. 3

The funnel plot of publication bias and standard error associated with Hedges’s g precision in the data of the twelve studies included in the meta-analysis. Black dots indicate inputted missing studies needed for the symmetrical plot