Differences in structural and functional default mode network connectivity in amyloid positive mild cognitive impairment: a longitudinal study.

Abstract

Default mode network (DMN) has emerged as a potential biomarker of Alzheimer's disease (AD); however, it is not clear whether it can differentiate amnestic mild cognitive impairment with altered amyloid (aMCI-Aβ +) who will evolve to AD. We evaluated if structural and functional connectivity (FC), hippocampal volumes (HV), and cerebrospinal fluid biomarkers (CSF-Aβ42, p-Tau, and t-Tau) can differentiate aMCI-Aβ + converters from non-converters. Forty-eight individuals (18 normal controls and 30 aMCI subjects in the AD continuum - with altered Aβ42 in the CSF) were followed up for an average of 13months. We used MultiAtlas, UF2C, and Freesurfer software to evaluate diffusion tensor imaging, FC, and HV, respectively, INNOTEST® kits to measure CSF proteins, and neuropsychological tests. Besides, we performed different MANOVAs with further univariate analyses to differentiate groups. During follow-up, 8/30 aMCI-Aβ + converted (26.6%) to AD dementia. There were no differences in multivariate analysis between groups in CSF biomarkers (p = 0.092) or at DMN functional connectivity (p = 0.814). aMCI-Aβ + converters had smaller right HV than controls (p = 0.013), and greater right cingulum parahippocampal bundle radial diffusivity than controls (p < 0.001) and non-converters (p = 0.036). In this exploratory study, structural, but not functional, DMN connectivity alterations may differentiate aMCI-Aβ + subjects who converted to AD dementia.