On the Reliability of the EEG Microstate Approach

Abstract

EEG microstates represent functional brain networks observable in resting EEG recordings that remain stable for 40-120ms before rapidly switching into another network. It is assumed that microstate characteristics (i.e., durations, occurrences, percentage coverage, and transitions) may serve as neural markers of mental and neurological disorders and psychosocial traits. However, robust data on their retest-reliability are needed to provide the basis for this assumption. Furthermore, researchers currently use different methodological approaches that need to be compared regarding their consistency and suitability to produce reliable results. Based on an extensive dataset largely representative of western societies (2 days with two resting EEG measures each; day one: n = 583; day two: n = 542) we found good to excellent short-term retest-reliability of microstate durations, occurrences, and coverages (average ICCs = 0.874-0.920). There was good overall long-term retest-reliability of these microstate characteristics (average ICCs = 0.671-0.852), even when the interval between measures was longer than half a year, supporting the longstanding notion that microstate durations, occurrences, and coverages represent stable neural traits. Findings were robust across different EEG systems (64 vs. 30 electrodes), recording lengths (3 vs. 2 min), and cognitive states (before vs. after experiment). However, we found poor retest-reliability of transitions. There was good to excellent consistency of microstate characteristics across clustering procedures (except for transitions), and both procedures produced reliable results. Grand-mean fitting yielded more reliable results compared to individual fitting. Overall, these findings provide robust evidence for the reliability of the microstate approach.

Keywords: Clustering; EEG microstates; Fitting; Retest reliability; Stability.

Fig. 1

Number of newly published studies per year on Scopus including the term “EEG microstates”

Fig. 2

Average short- and long-term retest-reliability of microstate characteristics ***p < .001, **p < .010, *p < .050, †p < .10. y-axis: Intraclass correlation coefficient (ICC) scale ranging from zero to one. x-axis: Microstate characteristics (Dur durations, Occ occurrences, Cov coverages, Trans transitions). Legend: k-means = k-means clustering, AAHC atomize and agglomerate hierarchical clustering, GM fitting = GM fitting procedure, Ind fitting = Ind fitting procedure. Top: Average ICCs of microstate characteristics across microstate types (A, B, C, C′, D) showing their short-term retest-reliability on day one (top left; N = 583) and day two (top right; N = 542). Bottom: Average ICCs of microstate characteristics across microstate types (A, B, C, C′, D) showing their long-term retest-reliability using pre-measures (bottom left; N = 525) and post-measures (bottom right; N = 525). Stars (and crosses) indicate significant (and marginally significant) differences between average ICCs obtained from different methodologies as shown by z-tests. We analyzed differences of average ICCs between clustering procedures (k-means/GM fitting vs AAHC/GM fitting and k-means/Ind fitting vs AAHC/Ind fitting) and fitting procedures (k-means/GM fitting vs k-means/Ind fitting and AAHC/GM fitting vs AAHC/Ind fitting)

Fig. 3

Average long-term retest-reliability of microstate characteristics in groups with different intervals between measures n = 525. y-axis: Intraclass correlation coefficient (ICC) scale ranging from zero to one. x-axis: Microstate characteristics (Dur durations, Occ occurrences, Cov coverages, Trans transitions). pre = pre-measures, post = post-measures, k-means = k-means clustering, AAHC atomize and agglomerate hierarchical clustering; Legend: Group 1 = interval of 1–7 days, Group 2: interval of 8–30 days, Group 3: interval of 31–90 days, Group 4: interval of 91–180 days, Group 5: interval of 181 days and more. Top: Average ICCs of microstate characteristics across types for each group, showing their long-term retest-reliability for pre-measures and k-means clustering (top left) and post-measures and k-means clustering (top right). Bottom: Average ICCs of microstate characteristics across types showing their long-term retest-reliability for pre-measures and AAHC (top left) and post-measures and AAHC (top right; see Table S15 in the supplementary material for information on group differences between average ICCs as indicated by z-tests). Notably, there was no systematic decrease of the retest-reliability with increasing intervals between day one and day two across all four conditions