Resting-state EEG microstates link neural dynamics to fluid intelligence in mild cognitive impairment

Background: Resting-state EEG microstates offer millisecond-scale markers of large-scale network dynamics that may capture early cognitive dysfunction in Mild Cognitive Impairment (MCI). We investigated whether microstate characteristics relate to global cognition and fluid intelligence in MCI. Methods: Sixty adults (30 MCI; 30 healthy controls, HC) were evaluated using the Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), and Raven's Progressive Matrices (RPM), and the Benton Facial Recognition Test (BFRT). Resting-state, eyes-closed EEG was recorded, and microstate Global Explained Variance (GEV) was calculated by mapping study specific topographies to the seven canonical classes (A-G) of the Metamaps2023 templates. In addition to GEV, temporal parameters (duration, occurrence, and coverage) were analyzed. Group differences and age/education/sex-adjusted correlations were assessed using a unified non-parametric framework with FDR correction. Results: = 0.007), while other classes showed no significant differences. Temporal analysis revealed that microstates B and C exhibited significantly increased occurrence and coverage in the MCI group, whereas mean duration did not differ between groups. Correlation analysis demonstrated that increased GEV of microstates B and C was associated with poorer performance on the MoCA and RPM, while microstates A and G showed positive correlations with MMSE scores. These associations remained robust after controlling demographic covariates. Conclusion: MCI is characterized by an increased contribution of microstates B and C, which are linked to impairments in fluid reasoning and executive control. These findings suggest that microstate dynamics serve as sensitive markers of early network disorganization in the predementia stage, reflecting a loss of neural flexibility that mirrors cognitive decline.