An outcome-blind EEG microstate decomposition in comatose cardiac arrest survivors identified a frozen microstate, M2, whose global explained variance was reproducibly associated with neurological outcome, but modest standalone discrimination suggests it complements rather than replaces established prognostic markers.
M2 global explained variance (GEV) was higher in comatose cardiac arrest survivors with favorable neurological outcome than in those with unfavorable outcome.
M2 GEV was inversely associated with CPC severity.
Standalone discrimination of neurological outcome by M2 GEV was modest.
M2 GEV remained independently associated with favorable outcome after adjustment for age, sex, and shockable rhythm.
The association persisted after additionally adjusting for ROSC time in a smaller complete-case cohort.
M2 GEV showed a graded association with outcome across quartiles without evidence of nonlinearity.
The M2 GEV association was robust to multiple sensitivity and resampling analyses.
Conventional temporal microstate parameters of M2 (coverage, mean duration, occurrence) were not significantly associated with outcome after multiplicity correction.
The microstate pipeline was outcome-blind and prespecified, using a frozen K = 4 solution backfitted to individual patients.
The author interprets M2 GEV as a marker of residual large-scale spatial organization that complements established clinical and background EEG markers.
This study re-analyzed EEG recordings from 567 comatose survivors of cardiac arrest in a large public database. Instead of looking at traditional EEG features like signal strength or suppression, it used 'microstates'—brief, recurring patterns of electrical activity across the scalp—identified without knowing patient outcomes. One of four patterns, called M2, accounted for more of the brain's overall electrical activity in patients who later had good neurological recovery than in those who did not. The link between M2 and outcome held up after accounting for age, sex, and whether the heart rhythm was shockable, and after many statistical stress tests. Patients with higher M2 values had higher odds of a favorable outcome, and the relationship was graded. However, the effect was modest: by itself, M2 distinguished good from poor outcomes only slightly better than chance (AUC 0.59). How often M2 appeared or how long it lasted was not meaningfully related to outcome. This research suggests that the spatial organization of brain activity after cardiac arrest carries some additional information about recovery. It does not suggest that M2 could replace existing prognostic tools; rather, it may serve as a complementary piece of information alongside clinical and standard EEG findings. Because this was a secondary analysis of existing data, the findings would need confirmation in further studies.
Machado C. (2026). Outcome-Blind EEG Microstate Topography After Cardiac Arrest: Spatial Expression of a Lateralized Microstate and Association with Neurological Outcome.. Brain topography. https://doi.org/10.1007/s10548-026-01267-7