Functional Connectome Analysis in Mild Cognitive Impairment: Comparing Alzheimer's Disease Continuum and Suspected Non-Alzheimer Pathology.

Abstract

Introduction: Research in brain resting-state functional connectivity (FC) analysis in mild cognitive impairment (MCI) has conflicting results. This work intends to find differences in resting-state FC of 2 groups of MCI subjects due to Alzheimer's disease (MCI-AD) continuum or to suspected non-Alzheimer pathology (MCI-SNAP). Materials and Methods: Ninety-two subjects older than 55 years were enrolled. MCI and controls were grouped using clinical dementia rating and neuropsychological data. Cerebrospinal fluid biomarkers were collected from MCI subjects, resulting in 32 MCI-AD, 25 MCI-SNAP, and 35 controls. A region of interest (ROI)-to-ROI analysis was carried out looking at inter- and intranetwork interactions selecting the following networks: default mode network (DMN), salience network (SN), visuospatial network (VN), and executive network. Pearson correlation coefficients, converted to Z-scores, were compared by T-tests with alpha set to 0.05, and false discovery rate corrected. Results: Groups were similar in age, education, and demographic measures, there were no differences in neuropsychological data between the MCI groups. The ROI-to-ROI analysis of MCI-AD versus MCI-SNAP showed no differences. MCI-AD versus controls showed decreased FC between ROIs of the SN and between ROIs from SN and VN. MCI-SNAP versus controls showed increased FC between an ROI of DMN and VN. Discussion: SN, DMN, and VN are multimodal networks with high value/high cost and may be more vulnerable to AD pathogenic processes. SN and VN were affected in the MCI-AD group, with maintained anticorrelation between DMN and VN. This may indicate subthreshold DMN dysfunction. The result in MCI-SNAP, although discrete, reflects a rearrangement of brain FC, as DMN and VN are expected to be anticorrelated. More research is necessary to confirm these findings.