Benchmarking foundation models for resting-state functional magnetic resonance imaging time series in cognitive aging research
- 1Stanford University
Abstract
Foundation models trained on large neuroimaging datasets could help smaller studies investigate cognition from brain recordings. However, whether the foundation model-derived representations improve assessment of current cognition, prediction of follow-up cognition, or characterization of cognitive change remains unclear. We evaluated nine configurations from five model families that process regional resting-state functional magnetic resonance imaging time series, using 3,103 scans from 1,653 participants across two observational and five intervention aging cohorts. Primary frozen representations showed no consistent superiority over conventional functional connectivity for concurrent cognitive associations or diagnostic classification. Changing the readout improved selected concurrent cognitive associations without updating model weights, indicating that performance depended partly on how representations were extracted. Additional post-training and embedding alignment yielded gains in selected settings but did not consistently improve performance across outcomes. In the primary longitudinal analyses, imaging representations did not establish added value beyond baseline clinical measures. These findings show that readout choice influences concurrent cognitive associations, while the added value of these representations for longitudinal outcomes remains unestablished.
Related
- Effects and mechanisms of exercise timing on cognitive function in older type 2 diabetes mellitus patients with mild cognitive impairment: A magnetic resonance imaging-based randomized controlled trial protocol
- Decision benchmarking reveals when protein foundation models add experimental value
- Benchmarking single-cell foundation models reveals an advantage in the low-data regime for cell-type and disease-state prediction
- Benchmarking single-cell foundation models for aging biology
- Harmonised benchmarking of foundation models for single-cell and spatial transcriptomics reveals context-dependent generalisation
BibTeX
@misc{liu2026benchmarking,
title = {{Benchmarking foundation models for resting-state functional magnetic resonance imaging time series in cognitive aging research}},
author = {Y. M. Liu and M. Ai and K. Zhao and Y. Zhang and F. V. Lin},
howpublished = {bioRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.05.764644},
url = {https://www.biorxiv.org/content/10.64898/2026.11.05.764644v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Benchmarking foundation models for resting-state functional magnetic resonance imaging time series in cognitive aging research
AU - Y. M. Liu
AU - M. Ai
AU - K. Zhao
AU - Y. Zhang
AU - F. V. Lin
PY - 2026
DA - 2026/11/05
PB - bioRxiv
DO - 10.64898/2026.11.05.764644
UR - https://www.biorxiv.org/content/10.64898/2026.11.05.764644v1
AB - Foundation models trained on large neuroimaging datasets could help smaller studies investigate cognition from brain recordings. However, whether the foundation model-derived representations improve assessment of current cognition, prediction of follow-up cognition, or characterization of cognitive change remains unclear. We evaluated nine configurations from five model families that process regional resting-state functional magnetic resonance imaging time series, using 3,103 scans from 1,653 participants across two observational and five intervention aging cohorts. Primary frozen representations showed no consistent superiority over conventional functional connectivity for concurrent cognitive associations or diagnostic classification. Changing the readout improved selected concurrent cognitive associations without updating model weights, indicating that performance depended partly on how representations were extracted. Additional post-training and embedding alignment yielded gains in selected settings but did not consistently improve performance across outcomes. In the primary longitudinal analyses, imaging representations did not establish added value beyond baseline clinical measures. These findings show that readout choice influences concurrent cognitive associations, while the added value of these representations for longitudinal outcomes remains unestablished.
ER -