Back

Patient-specific functional brain architecture explains cortical patterns of tau PET in Alzheimer's disease

Behjat, H. H.; Vogel, J. W.; Strandberg, O.; Spotorno, N.; Rittmo, J.; Collij, L. E.; Pichet Binette, A.; Xiao, Y.; Chaggar, P.; van Westen, D.; Stomrud, E.; Palmqvist, S.; Mattsson-Carlgren, N.; Van De Ville, D.; Smith, R.; Hansson, O.; Ossenkoppele, R.

2025-10-03 neuroscience
10.1101/2025.10.02.679969 bioRxiv
Show abstract

The spatial distribution of tau pathology, the core driver of neurodegeneration in Alzheimers disease (AD), varies markedly across individuals. While tau is thought to spread along brain networks, the role of inter-individual variability in shaping these patterns remains underexplored. Using resting-state fMRI and tau-PET from 805 participants across the AD continuum, we studied whether subject-specific functional connectivity (FC) profiles enhance the characterization of tau deposition patterns. A hybrid approach integrating individual and group-average FC outperformed both alone, particularly in symptomatic individuals and at finer spatial resolutions, the latter underscoring a critical but often overlooked role of spatial scale. Individualized FC also better captured individual tau topographies than canonical tau-PET maps derived from cohort-level data. These effects were specific to tau, and not seen for {beta}-amyloid, and their predictive power increased with spatial granularity. Furthermore, baseline FC also predicted future tau accumulation at the individual level, supporting its prognostic value. Together, these findings provide strong evidence that individual functional brain architecture shapes tau propagation in humans, supporting the network spread hypothesis by showing that variability in connectivity translates into heterogeneity in tau distribution. This work advances biological understanding of tau propagation in AD, highlighting functional connectivity as a mechanistic substrate that supports prognostic assessment of tau trajectories.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.