Electroencephalography (EEG) microstate analysis represents a powerful computational approach for investigating the temporal dynamics of large-scale cortical networks in neurodegenerative diseases. Microstates are brief periods (typically 50-120 milliseconds) of quasi-stable scalp potential distributions that reflect coordinated activity of distinct neuronal networksMusaeus CS 2024, Microstate changes associated with Alzheimer. The systematic review by Musaeus et al. (2024) examining EEG microstates in Alzheimer's dementia (AD) and mild cognitive impairment (MCI) found that pooled evidence suggests prolonged microstate A and B duration as the most reproducible alterations in AD, with reduced microstate D duration emerging as a modest finding in MCIMusaeus CS 2024, EEG microstates in Alzheimer.
EEG microstate analysis offers several advantages over traditional spectral EEG approaches. While conventional analysis focuses on frequency bands (delta, theta, alpha, beta, gamma), microstate analysis captures the spatiotemporal organization of brain activity, providing information about the sequential organization of cortical networksKhomenko E 2020, EEG microstates as a biomarker for early detection of Alzheimer. This approach has been increasingly applied to understand the network-level disruptions that underlie cognitive decline in AD and MCI.
The microstate concept originated from the work of Walter Diesch and colleagues in the 1990s, who proposed that the EEG signal could be parsed into a sequence of quasi-stable brain statesDiesch E 1995, EEG microstates in cognitive neuroscience. Each microstate represents a momentary configuration of the electromagnetic field generated by synchronized postsynaptic potentials across large neuronal populations. The spatial configuration of these potentials remains relatively stable before transitioning to a different configuration.
The analysis typically employs a clustering algorithm (most commonly k-means or modified k-means) applied to the EEG topographic maps at latency points of maximum variance. This approach identifies a finite set of microstate classes that account for the majority of variance in the scalp potential distributions. Research has consistently identified four main microstate classes (A, B, C, D) that explain 60-80% of the global variance in resting-state EEGBrodbeck C 1992, Topographic analysis of spontaneous EEG.
The four canonical microstate classes have been associated with distinct functional brain networks:
These assignments remain somewhat controversial, as the functional interpretation of microstates varies across studies and populations[1]. However, the consistency of four microstate classes across multiple datasets suggests a fundamental organization of spontaneous brain activity into discrete network configurations.
The primary parameters extracted from microstate analysis include:
These parameters provide complementary information about brain network dynamics. Duration reflects the stability of a network configuration, occurrence reflects how frequently a network is engaged, and coverage reflects the overall time spent in each network state.
The systematic review by Musaeus et al. identified significant increases in microstate A duration (g = 0.41, 95% CI [0.10, 0.72]) and microstate B duration (g = 0.48, 95% CI [0.23, 0.73]) in AD patients compared to healthy controlsMusaeus CS 2024, EEG microstates in Alzheimer. These findings suggest that AD is associated with prolonged periods of stability in specific network configurations.
The increased duration of microstates A and B may reflect impaired switching between brain states, potentially due to neurodegeneration in the temporal and visual networks. The temporal lobe is particularly vulnerable in AD, with early accumulation of amyloid plaques and neurofibrillary tangles in entorhinal cortex and hippocampus. Disruption of temporal-parietal networks may explain the prolonged microstate A configurations associated with auditory processing.
Similarly, microstate B alterations may reflect visual pathway dysfunction in AD. Visual processing deficits are common in AD and may precede memory impairment in some patients. The visual network alterations could be related to amyloid deposition in visual association areas or disrupted cholinergic modulation of visual processing.
Several studies have reported additional microstate alterations in AD:
Recent work by Yang et al. (2024) demonstrated that resting-state EEG microstate features can classify AD patients with high accuracy, with microstate duration and coverage being the most discriminative parametersYang X 2024, Resting-state EEG microstate features for Alzheimer. Yan et al. (2024) found that abnormal EEG microstates serve as predictors of beta-amyloid deposition degree, suggesting a direct link between microstate alterations and AD pathologyYan Y 2024, Abnormal EEG microstates in Alzheimer.
The systematic review noted substantial heterogeneity for several outcomes, and possible small-study effects were identifiedMusaeus CS 2024, EEG microstates in Alzheimer. This heterogeneity likely reflects methodological differences across studies, including:
The authors concluded that current evidence is best interpreted as hypothesis-generating pending standardized, longitudinal, and multimodal studies.
In MCI patients, the systematic review found a significant decrease in microstate D duration (g = -0.26, 95% CI [-0.48, -0.04]) compared to healthy controlsMusaeus CS 2024, EEG microstates in Alzheimer. This finding is particularly interesting because MCI often represents a transitional stage between normal aging and AD, and microstate D alterations may represent an early biomarker of impending dementia.
Microstate D is associated with the dorsal attention network, which is involved in top-down attention and working memory. The reduced duration of this microstate in MCI may reflect early dysfunction in attention networks that precedes the more widespread network disruption seen in full-blown AD.
The review also found increased microstate A occurrence rate in MCI (g = 0.40, 95% CI [0.07, 0.74])Musaeus CS 2024, EEG microstates in Alzheimer. This suggests that individuals with MCI more frequently engage temporal network configurations, potentially as a compensatory mechanism or as an early sign of temporal network dysfunction.
Lian et al. (2021) confirmed these findings in a study showing altered EEG microstate dynamics in MCI, with distinct patterns compared to both AD and healthy controlsLian H 2021, Altered EEG microstate dynamics in mild cognitive impairment and Alzheimer. The study demonstrated that microstate parameters can differentiate MCI from AD, potentially aiding in disease staging.
These findings support the potential of EEG microstate analysis as a biomarker for early detection. The alterations in MCI are distinct from those in AD, suggesting that microstate parameters may help differentiate between disease stages. This could have important implications for:
EEG microstate analysis offers several advantages as a diagnostic biomarker for AD and MCI:
Despite these advantages, several limitations currently constrain clinical adoption:
Recent work by Simmatis et al. (2025) proposed an automated pipeline for detecting and monitoring AD progression using EEG biomarkers, addressing some of these limitations through standardized analysisSimmatis LE 2025, EEG biomarkers for Alzheimer.
Future clinical application likely requires integration of microstate analysis with other biomarkers. Combining EEG with:
Multimodal approaches may improve diagnostic accuracy and provide more complete characterization of disease state.
EEG microstate analysis may serve as a tool for monitoring therapeutic response in AD and MCI. Several potential applications include:
EEG microstate analysis enables real-time visualization of brain network states, making it suitable for neurofeedback applications. Neurofeedback aims to train patients to modulate their brain activity in desired directions. For AD and MCI, neurofeedback could potentially:
Preliminary studies have explored EEG neurofeedback in AD with mixed results, but microstate-targeted approaches have not yet been thoroughly investigated.
The microstate alterations in AD may inform development of novel therapeutic approaches:
Hanoglu et al. (2022) explored the therapeutic role of repetitive transcranial magnetic stimulation in AD and PD, finding EEG microstate correlates with clinical outcomesHanoglu L 2022, Therapeutic role of repetitive transcranial magnetic stimulation in Alzheimer.
The field requires standardization in several areas:
Longitudinal studies are needed to establish:
Future research should integrate microstate analysis with:
EEG microstate analysis provides a window into the temporal organization of large-scale cortical networks in AD and MCI. The systematic review evidence suggests that prolonged microstate A and B duration represents the most reproducible alteration in AD, while reduced microstate D duration in MCI may serve as an early biomarker of network dysfunction.
While methodological heterogeneity currently limits interpretation, these findings support the hypothesis that AD involves impaired network state switching, potentially reflecting neurodegeneration in specific cortical systems. The non-invasive, inexpensive nature of EEG makes microstate analysis attractive for clinical applications, but standardization and validation are needed before widespread clinical adoption.
Future research should focus on standardized protocols, longitudinal studies, and multimodal integration to establish the full potential of EEG microstates as biomarkers and therapeutic monitoring tools in neurodegenerative disease.
Custo A, Van De Ville D, Michel C, De Lucia M. Aging effects on EEG microstates: a 5-year longitudinal study. Neurobiology of Aging. 2014. ↩︎