Organizational principles of the cerebral cortex predict symptoms progression in the Alzheimer’s disease spectrum

Functional gradients offer a novel, synthetic and interpretable view of functional connectivity. By integrating gradients, graph theory, morpho-volumetric indices, amyloid status, and regional tau burden, we aim to understand how cognitive decline relates to functional and structural brain changes. We analyzed 279 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database with available functional, structural, and proteomic data at baseline and at 2-year follow-up, alongside detailed assessments of verbal memory abilities (encoding, retrieval and recall). Results at baseline revealed distinct patterns based on amyloid status and clinical severity. In individuals with positive amyloid status, increased functional dispersion of the Default Mode Network (DMN) correlated with poorer memory, while in individuals with negative status higher dispersion was associated with better performance. Notably, middle temporal regions exhibited connectivity profiles opposite to other DMN areas. We also observed that regions where gradients' dispersion predicted impaired memory performance showed preserved structural integrity but had a more widespread connectivity profile. Baseline dispersion predicted future memory decline, showing consistent patterns across positive and negative amyloid groups-higher dispersion was linked to faster memory decline, except in tau-rich regions like the temporal cortex, where it was advantageous. Furthermore, higher dispersion was associated with future tau buildup and morphological alterations. This study extends a previous investigation and evaluates the utility of a multimodal framework in tracking disrupted DMN activity as an early marker of reduced coherence in communication among brain regions, possibly having a predictive value for future progression of cognitive decline.