Early alterations in the cognitive connectome precedes Mild Cognitive Impairment

INTRODUCTION: Focusing on topographic statistic dependencies, the cognitive connectome framework provides a more sensitive detection of cognitive alterations. This study examines connectome patterns in predementia stages. METHODS: A neuropsychological battery measured cognitively healthy older adults (N = 822) at two timepoints, comparing stable controls with Mild Cognitive Impairment (MCI) converts (n = 617; n = 116). Networks were estimated as regularized Gaussian graphical models (EBICglasso, 20 nodes, 190 edges), followed by Frobenius distance with permutations for global differences. RESULTS: Significant differences between groups emerged at T0 (50/190 nodes, p < .005), and increasing at T1 (78/190, p < .005). Differences were predominantly intermodular (68% baseline; 76% longitudinal), involving phonological/semantic fluency, episodic memory, processing speed, and working memory. DISCUSSION: MCI conversion is preceded by architectural cognitive reorganization and altered cross-domain integration. The cognitive connectome provides a complementary early-detection tool for individual at risk for MCI in preclinical stages.