Longitudinal sensory-evoked fMRI reveals early hyperactivity, network hypersynchrony, and modular disruption in a preclinical model of Alzheimer’s disease

Alzheimer’s disease (AD) unfolds over a prolonged asymptomatic phase in which neural and network dysfunction precedes overt cognitive decline. However, sensitive noninvasive biomarkers that capture both local circuit abnormalities and large-scale network reorganization remain limited. Here, we performed longitudinal functional magnetic resonance imaging (fMRI) in the Swedish-Dutch-Iowa (Tg-SwDI) mouse model of AD and combined visual- and olfactory-evoked fMRI with temporal, hierarchical, clustering, and resting-state connectivity analyses. Visual stimulation revealed widespread hyperactivity at 2 months, particularly in hippocampal, prefrontal, thalamic, and higher-order visual regions, whereas olfactory abnormalities were more spatially restricted and most prominent at 3 months. These modality-specific changes were accompanied by altered response dynamics, increased temporal synchrony, and reduced resting-state modularity and network segregation at 3 months, before marked spatial working-memory impairment at 6 months. By 6 months, several prefrontal and thalamic regions showed reduced responsiveness. Together, these findings support a longitudinal framework in which early sensory-evoked hyperactivity is followed by a transitional state characterized by emerging hypersynchrony, abnormal temporal coordination, and reduced network segregation, and subsequently by regional hypoactivity and marked cognitive impairment. These convergent alterations suggest progressive destabilization of large-scale network organization during disease progression. Sensory-evoked fMRI may therefore provide a behavior-independent, systems-level approach for detecting early functional vulnerability and identifying candidate biomarkers of preclinical AD.