Brain aging and Alzheimer’s disease alter connections between brain regions in distinct and independent ways
Brain aging and Alzheimer’s disease alter connections between brain regions in distinct and independent ways
Researchers at Lund University tested this finding using brain scans from 973 Swedish participants in the BioFINDER-2 cohort, then independently in 129 American participants in the ADNI initiative. Changes associated with aging and with the accumulation of Alzheimer’s marker proteins mapped onto two distinct axes of brain organization. The findings were reproduced in both samples and in follow-up scans of the same participants years later. The study was published in Nature Neuroscience on October 1, 2026.
For decades, measurements of brain connectivity in aging and Alzheimer’s disease produced conflicting results: the same region could appear more connected in one analysis and less connected in another, leaving it uncertain whether one process or two were involved. Earlier work from the same laboratory offered a clue. In 2020, its head, Jacob Vogel, showed in humans that tau protein in Alzheimer’s disease spreads through the brain along connections between neurons, following them like roads. The researchers reasoned that if the pathology travels along these routes, it should leave a distinct pattern in the brain’s connections that can be distinguished from the effects of normal aging.
To separate the two processes mathematically, the authors used a continuous pathology score based on tau and amyloid proteins in place of diagnostic categories such as cognitively healthy, mild cognitive impairment, or dementia. These categories conflate disease severity with a person’s cognitive reserve. For each of 1000 cortical regions, they calculated “nodal similarity,” a measure of how closely its pattern of connections resembles those of all other brain regions, and constructed separate maps of age-related and pathology-related effects. They compared both maps with two principal axes of brain organization: one runs from regions involved in basic sensory processing to those involved in complex, abstract processing (the sensory-association axis), while the other separates regions that store information and memories from those that direct attention and decision-making (the representation-execution axis).
The pathology map closely matched the sensory-association axis, while the age map showed an inverse correspondence with the representation-execution axis. The result was replicated in the independent ADNI sample. This is consistent with established symptom patterns: Alzheimer’s disease primarily affects memory, while normal aging affects the ability to plan, sustain attention, and make decisions. Cognitive tests supported the same pattern: memory was associated with the first axis, and these other abilities with the second.
The pathology signature is already present early in the disease and does not increase with disease severity. “We were surprised to find that the characteristic pattern of changes in brain function was already apparent in people with low levels of Alzheimer’s pathology who remained cognitively healthy,” says the study’s first author, Jonathan Rittmo. Among these participants, alignment with the sensory-association axis peaked at a moderate level of pathology accumulation. In participants who had already received a diagnosis, that alignment disappeared: the same axis began to reflect the severity of current cognitive symptoms, independently of tau and amyloid protein levels.
The BioFINDER cohort is led by Oskar Hansson, whose laboratory in Lund brought the p-tau217 blood test into clinical use. It is one of the few biomarkers of early Alzheimer’s disease actually used in practice.
For now, these findings apply to large groups of people. The authors identify the next step as testing whether the signature predicts future cognitive decline in an individual patient. Vogel has already suggested a possible application if the signature proves to reflect harmful strain on the brain’s systems: “This could suggest ways to intervene, for example through noninvasive brain stimulation.”