The intrinsic connectivity between the default mode and dorsal attention networks is an independent fMRI biomarker of Alzheimer's disease pathology burden

The mechanism of neurocognitive failure in Alzheimer's disease remains obscure. While the mainstream hypothesis in the field posits that brain tau pathology is the only process that drives cognitive decline in AD, other complementary mechanisms link vascular brain lesions with beta-amyloid pathology as an important factor leading to neurodegeneration. Recently, it was also proposed that the brain's network's functional imbalance could primarily drive cognitive decline in neurodegenerative diseases. Here, we investigated whether the anticorrelation between the default mode (DMN) and dorsal attention networks (DAN) reveals different pathology burdens in the AD spectrum. We grouped individuals based on their PET amyloid and cognitive status. Using cross-validated regression models, we investigated whether cognitive impairment can be predicted based on rs-fMRI DMN-DAN anticorrelation. We found that the DMN-DAN anticorrelation differentiates between pathology burdens in AD, as quantified by PET amyloid imaging and cognitive performance. We found that an attenuated DMN-DAN anticorrelation predicted cognitive decline, which was controlled by sex, age, education, and brain tau pathology. Education level, measuring cognitive reserve, did not modulate the association between DMN-DAN anticorrelation and cognitive decline. We demonstrate that the attenuation of the anticorrelation between DMN and DAN is associated with a mechanism of cognitive dysfunction independent of tau pathology and proxies of resilience to cognitive decline or cognitive reserve. Our results also suggest the existence of an alternative mechanism of neurocognitive breakdown independent of advanced medial temporal cortex pathology and protective factors of cognitive decline, such as cognitive reserve.