Alzheimer's rewires brain communication along its own path before the cortex shrinks

Alzheimer's pathology builds up mostly in old age. Any older brain in the scanner therefore carries both influences at once, and telling healthy aging apart from early network failure has been hard ^[7]. Jacob Vogel's group at Lund dealt with that through the cohort. It included cognitively unimpaired young adults as well as older adults and people across the range of Alzheimer's pathology, all studied with resting-state functional MRI ^[6]. With that mix, age effects can be measured in brains without pathology and then compared with pathology effects in older brains ^[6].

The changes did not stay in single regions. According to the Lund University account of the study, they moved together across the whole brain as coordinated reorganizations along its main functional gradients ^[8][11]. The timing is what caught the researchers' attention. "We already know that blood markers such as p-tau217 can reveal signs of Alzheimer's disease before cognitive symptoms appear. What surprised us was that the distinctive pattern of changes in brain communication was already apparent in people with low levels of Alzheimer's pathology who were still cognitively unimpaired," said first author Jonathan Rittmo, a doctoral researcher at Lund University ^[5].

That early group was defined by its low biomarker levels ^[3]. Protein measures still decided who counted as early-stage before anyone looked at the wiring.

The comparison with plaques and tangles has the most bearing on drug trials, and it was made in one specific group. It was run in cognitively impaired participants, and it set whole-brain connectivity against the regional accumulation of amyloid-beta and tau ^[4]. A closer correlation with test scores at that stage says connectivity reflects how a patient is functioning now. It does not show that the wiring change drives the memory loss. Nor does it show that connectivity predicts which unimpaired people will go on to decline. The release does not report correlation coefficients, and it does not say whether the trajectories come from repeat scans of the same people.

I think the finding supports adding a connectivity readout alongside protein measures when a trial asks whether patients function better. The case for finding patients by connectivity instead of protein burden is weaker. As Rittmo noted, blood markers such as p-tau217 already flag the disease before symptoms appear ^[5]. Both judgements depend on the link to cognition holding up in a second cohort and over time.