Aberrant default mode network in amnestic mild cognitive impairment: a meta-analysis of independent component analysis studies

Independent component analysis (ICA) is one of the most popular and valid methods to investigate the default mode network (DMN), an intrinsic network which attracts particular attention in amnestic mild cognitive impairment (aMCI). However, previous studies present inconsistent results regarding the topographical organization of the DMN in aMCI. Therefore, we conducted a quantitative, voxel-wise meta-analysis of resting-state ICA studies using Seed-based d Mapping to establish the most consistent pattern of DMN functional connectivity alterations in aMCI. Twenty studies, comprising 23 independent datasets involving 535 patients and 586 healthy controls, met the inclusion criteria. Patients with aMCI exhibited reliably lower DMN functional connectivity than the healthy controls in the bilateral precuneus/posterior cingulate cortices and medial temporal lobes, which are implicated in episodic memory deficits. Moreover, an exploratory meta-regression analysis revealed that greater severity of global cognitive impairment in the patient groups was associated with stronger functional connectivity in the bilateral medial frontal cortices (including the anterior cingulate cortices), left angular gyrus, and right temporal pole extending to the middle temporal gyrus, likely reflecting a compensatory mechanism for maintaining cognitive efficiency. This meta-analysis identifies a consistent pattern of aberrant DMN functional connectivity in aMCI, which facilitates understanding of the neurobiological substrates of this disease.