Counterfactual aging in individual digital twin brains uncovers a latent dimension of resilience to aging

Cognitive reserve concerns why the same aging causes less decline in some individuals than others, a question that digital twin brains can address by aging each individual in silico. Here we present Brain-VRNN, a connectome-constrained variational recurrent neural network trained on longitudinal multitask functional MRI. Its regenerated brain signals reproduced four functional aging markers of unseen individuals (r = 0.93–0.96) and identified each person (accuracy 0.98). Without cognitive scores, its global latent learned a trait dimension stable within each person, which helped predict future cognition. Crucially, when every individual underwent the same counterfactual aging, those with a lower trait value declined more, yet recovered up to 1.6 times as much with a reduction in body mass index, a modifiable dementia risk factor. Aging single regions showed that different markers decline through different regions, with a lateral temporal–angular region shared by all. Brain-VRNN brings within-person simulation to cognitive reserve.

Authors

Institutions

Publication Details

Journal
bioRxiv (Cold Spring Harbor Laboratory)
Published
2026-11-18
DOI
https://doi.org/10.64898/2026.11.18.767485
Primary Topic
Functional Brain Connectivity Studies
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Counterfactual aging in individual digital twin brains uncovers a latent dimension of resilience to aging

Yuichi Yamashita, Hayato Idei, Yuta Takahashi, Hiroaki Tomita
1 links
bioRxiv (Cold Spring Harbor Laboratory)
Functional Brain Connectivity Studies
preprint

Counterfactual aging in individual digital twin brains uncovers a latent dimension of resilience to aging

Yuichi Yamashita, Hayato Idei, Yuta Takahashi, Hiroaki Tomita
preprint en

Abstract

Cognitive reserve concerns why the same aging causes less decline in some individuals than others, a question that digital twin brains can address by aging each individual in silico. Here we present Brain-VRNN, a connectome-constrained variational recurrent neural network trained on longitudinal multitask functional MRI. Its regenerated brain signals reproduced four functional aging markers of unseen individuals (r = 0.93–0.96) and identified each person (accuracy 0.98). Without cognitive scores, its global latent learned a trait dimension stable within each person, which helped predict future cognition. Crucially, when every individual underwent the same counterfactual aging, those with a lower trait value declined more, yet recovered up to 1.6 times as much with a reduction in body mass index, a modifiable dementia risk factor. Aging single regions showed that different markers decline through different regions, with a lateral temporal–angular region shared by all. Brain-VRNN brings within-person simulation to cognitive reserve.

bioRxiv (Cold Spring Harbor Laboratory)
Tohoku University (JP), National Center of Neurology and Psychiatry (JP)
Japan Agency for Medical Research and Development, National Center of Neurology and Psychiatry, Japan Society for the Promotion of Science, Core Research for Evolutional Science and Technology
Good health and well-being
Functional Brain Connectivity Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon