Researchers using FDG-PET brain scans have identified two opposing patterns of glucose metabolism in the brain's white matter that predict how quickly a person's cognition will decline in the following years, according to a study published March 19, 2026 in Nature Communications. The team, led by investigators at the Mayo Clinic, analyzed imaging and cognitive data from 3,142 participants across two long-running aging studies, the Mayo Clinic Study of Aging (MCSA) and the Alzheimer's Disease Neuroimaging Initiative (ADNI), with more than 15,000 follow-up visits combined.
Most prior FDG-PET research has focused on gray matter, where lower glucose uptake is a well-established marker of Alzheimer's disease. This study instead mapped white matter voxel by voxel and found two separate signatures: one in the corpus callosum and cingulum, where higher metabolism tracked with better cognition, and an opposite one in the corona radiata, where higher metabolism tracked with worse cognition. The authors describe the second pattern as a possible sign of the brain working harder to compensate as disease progresses. The corpus callosum and cingulum pattern weakened with age, amyloid buildup, the APOE-e4 gene variant, and white matter hyperintensities (changes often linked to small vessel disease), while the corona radiata pattern tended to rise with age and vascular risk factors but was blunted by amyloid.
In statistical models, adding both white-matter signatures to standard measures increased the variance in cognitive scores that could be explained from 43 percent to 46 percent in the MCSA cohort and from as low as 23 percent to 37 percent in ADNI, depending on which baseline measure was being improved upon. Longitudinally, people with higher metabolism in the corpus callosum and cingulum pattern and lower metabolism in the corona radiata pattern showed slower cognitive decline over time, a relationship that held after accounting for amyloid status and vascular risk.