Premium
Latent patterns of task‐related functional connectivity relate to hyperactivation and associative memory impairment in SCD + and MCI
Author(s) -
CorriveauLecavalier Nick,
Rajah Natasha,
Mellah Samira,
Belleville Sylvie
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.038002
Subject(s) - hippocampus , memory impairment , cognition , audiology , psychology , hippocampal formation , neuroscience , episodic memory , medicine
Background Hyperactivation is commonly observed in prodromal Alzheimer’s disease (AD). It is however unknown if it relates to functional network dysfunction, and how hyperactivation‐networks relate to cognitive impairment. We used seed‐based connectivity to assess differences in correlations between activation of hyperactive regions, whole‐brain task‐related activation, and memory performance in persons with subjective cognitive decline plus (SCD + ) and mild cognitive impairment (MCI). Method Data from 108 participants from the CIMA‐Q cohort were used in this study: 28 participants with SCD + which presented with memory complaint and worry in addition to small hippocampal volume and/or were APOE4 carriers, 26 participants with MCI and 54 were healthy controls (HC). Task‐related activation was measured while participants memorized 78 pictures and their location in a four‐position grid. Multivariate seed‐partial least squares (seed‐PLS) analysis was used to identify latent variables (LVs) that assess between‐group differences and similarities in the three‐way association between 1) activation in hyperactive regions of interest (ROIs: left hippocampus, right inferior temporal gyrus (riTG), left superior parietal lobule (lsPL)), 2) whole‐brain activation associated with associative memory encoding, and 3) subsequent associative memory performance. Significance of the seed‐PLS effects were assessed with permutation testing and boostraps. Result Three LVs were significant ( p < 0.05). LV1 (32.11 % of covariance explained) accounted for group similarities in the correlation between activity in all ROIs and activity in the fronto‐parietal and default mode networks. Interestingly, left hippocampus activity did not correlate in SCD + . LV2 (19.17 % of covariance explained) identified a network that was correlated with left hippocampal activity in all three groups; however, left hippocampal activity correlated with better memory in HC but with poorer memory in SCD + . LV3 (10.91 % of covariance explained) identified a network that was positively correlated to lsPL activity in SCD + , and correlated with seed activity in all ROIs and poorer memory performance in MCI. Conclusion Hyperactivation is associated with latent patterns of connectivity in prodromal AD, but different hyperactivation‐network associations relate to memory in SCD + and MCI. Hyperactivation could serve as an early marker of network dysfunction in AD, and reflect the incoming cognitive symptomatology associated with the disease.