Abstract / Summary
Abstract Objective Electroencephalography (EEG) is central to neurological assessment after cardiac arrest, yet conventional approaches predominantly characterize amplitude, frequency content, continuity, suppression, reactivity, or predefined pathological patterns. EEG microstates provide a complementary framework by describing quasi-stable configurations of the global scalp electric field. We investigated whether an outcome-blind microstate decomposition of EEG from comatose survivors of cardiac arrest identifies reproducible spatial states associated with neurological outcome. Methods We performed a secondary analysis of the I-CARE/PhysioNet database. After prespecified EEG eligibility screening, 567 subjects were included. Microstate reconstruction was performed outcome-blind using 19-channel EEG, a fixed 300-s analysis window, GFP-peak sampling, and polarity-invariant clustering across K=2–10. After model-order evaluation, the K=4 solution was frozen and backfitted at the patient level. Associations between microstate metrics and neurological outcome were subsequently evaluated using unadjusted, multivariable, dose-response, and resampling analyses. Results The primary outcome analysis included 567 patients: 209 with favorable and 358 with unfavorable neurological outcomes. M2 global explained variance (GEV) was higher in patients with favorable outcome (median 0.0852 vs. 0.0564; rank-biserial correlation=0.185; p=0.000243; false-discovery-rate [FDR]-adjusted p=0.000568 within the four-state GEV family). M2 GEV was inversely associated with CPC severity (Spearman rho = -0.144; p = 0.000583; FDR-adjusted p = 0.001419). After adjustment for age, sex, and shockable rhythm in the extended cohort without return-of-spontaneous-circulation (ROSC) time (N=535), each 1-SD increase in M2 GEV was associated with greater odds of favorable outcome (OR 1.30, 95% CI 1.07-1.58; p=0.0092). In the complete-case model additionally adjusted for ROSC time (N=292), the association remained significant (OR 1.47, 95% CI 1.10-1.97; p=0.0087). A graded association was observed across M2 GEV quartiles without evidence of nonlinearity. Results remained directionally stable under winsorization, exclusion of distributional extremes, heteroscedasticity-consistent standard errors, bootstrap resampling, leave-one-out analysis, and repeated leave-five-out perturbations. By contrast, M2 coverage, mean duration, and occurrence were not significantly associated with outcome after multiplicity correction. Conclusions An outcome-blind decomposition of EEG after cardiac arrest identified a frozen microstate, M2, whose GEV was reproducibly associated with neurological outcome. The prognostic information associated with M2 was related primarily to the degree to which this spatial configuration explained the global EEG field rather than to its conventional temporal parameters. EEG microstate architecture may provide a complementary description of residual large-scale electrophysiological organization after hypoxic-ischemic brain injury. External validation is required before clinical application.