Modeling the Switching behavior of Functional Connectivity Microstates (FCμstates) as a Novel Biomarker for Mild Cognitive Impairment

摘要: It is evident the need for designing and validating novel biomarkers detection of mild cognitive impairment (MCI). MCI patients have a high risk developing Alzheimers disease (AD) that reason introduction reliable significant clinical importance. Motivated by recent findings about rich information dynamic functional connectivity graphs (DFCGs) brain (dys)function, we introduced approach identifying based on magnetoencephalographic (MEG) resting state recordings. The activity different rhythms (delta, theta, alpha1, alpha2, beta1, beta2, gamma, gamma2) was first beamformed with linear constrained minimum norm variance in MEG data to determine ninety anatomical regions interest (ROIs). A graph (DFCG) then estimated using imaginary part phase lag value (iPLV) both intra-frequency coupling (8) also cross-frequency pairs (28). We analyzed DFCG profiles neuromagnetic recordings 18 Mild Cognitive Impairment (MCI) 20 healthy controls. followed our model dominant intrinsic mode (DICM) across sources temporal segments further leads construction an integrated (iDFCG). filtered statistically topologically every snapshot iDFCG data-driven approaches. Estimation normalized Laplacian transformation segment related eigenvalues created 2D map network metric time series (NMTSeigs). NMTSeigs preserves non-stationarity fluctuated synchronizability iDCFG each subject. Employing initial set elders patients, as training set, built overcomplete dictionary microstates (FCmstates). Afterward, tested whole procedure extra blind subjects external validation. succeeded classification accuracy dataset (85 %) which supports proposed Markovian modeling evolution states. adaptation appropriate neuroinformatic tools combine advanced signal processing neuroscience could manipulate properly time-resolved FC patterns revealing robust biomarker MCI.