Researchers found that dysconnectivity in the brain's default-mode network may predict future dementia, with an area under the curve of 0.82 and time to diagnosis (R = 0.53), outperforming traditional models based on brain structure.
A recent study investigated the predictive power of examining the brain's default-mode network (DMN) effective connectivity for early dementia detection. The study, published in Nature Mental Health, included 81 undiagnosed individuals who developed dementia within nine years post-imaging and 1,030 matched controls.
Researchers used spectral dynamic causal modeling on resting-state functional MRI (fMRI) data from the UK Biobank. Resting-state fMRI data were analyzed to estimate effective connectivity in the DMN. The study examined whether these connectivity changes could predict future dementia diagnosis and the time until diagnosis. Participants were matched on age, sex, ethnicity, handedness, and imaging center location.
The model accurately predicted dementia onset up to nine years before an official diagnosis was made, reported the investigators, with an accuracy of more than 80%.
The study also found strong associations between DMN dysconnectivity and major dementia risk factors—particularly, polygenic risk for Alzheimer’s disease and social isolation.
Effective connectivity mapping in the DMN was identified as a method for early dementia detection, which may help predict dementia at the individual level.
The study authors concluded, "We found that effective connectivity in the DMN can be used as a non-invasive population-based prediagnostic biomarker for predicting future dementia incidence. This biomarker, using [fMRI] data, is superior to using structural MRI data."
The authors reported no conflicts of interest.