Resting-state fMRI foundation models enable robust and generalizable latent neural target discovery in cognitive aging interventions

Abstract The benefits of interventions targeting cognitive aging vary substantially across individuals, largely owing to heterogeneity in aging-related comorbidities. It is necessary to robustly identify neural patterns underlying intervention response and test their generalizability across heterogeneous cohorts. Resting-state functional MRI (rsfMRI) offers a potential pathway, but relying on predefined summary features with conventional methods has limited capacity to capture both within-individual longitudinal variation and between-individual differences, particularly in small and heterogeneous studies. Recent rsfMRI foundation models pretrained on large observational cohorts present a promising alternative by learning transferable spatiotemporal representations from time-series signals. Yet their validity and generalizability in local intervention settings remain unclear. Here, we systematically evaluated rsfMRI foundation models using data from two independent randomized controlled trials of older adults with mild cognitive impairment, testing whether these models can robustly extract longitudinal brain representations that predict post-intervention changes in episodic memory across trials. Foundation models outperformed conventional machine learning and deep learning approaches across both trials. Clinically informed adaptation using an external Alzheimer’s disease cohort further improved performance and robustness to confounders (i.e., head motion, site, and intervention arm), with accuracy up to 82%. Multivariate decomposition of foundation model embeddings identified latent neural patterns associated with episodic memory change with cross-study consistency at baseline that became more spatially distributed at post-intervention. These findings show that rsfMRI foundation models can enable robust and generalizable identification of latent neural patterns linking longitudinal brain dynamics to individual intervention response, laying the foundation for precision-driven neural target discovery in cognitive aging research. Significance Statement Interventions for cognitive aging show highly variable outcomes across individuals, limiting their clinical effectiveness. This study introduces a foundation model-based approach to identify latent neural patterns underlying individual differences in intervention response using resting-state fMRI. By leveraging pretrained models and domain-adaptive fine-tuning, we demonstrate robust and generalizable prediction of cognitive improvement across independent trials. Our findings suggest that latent brain representations, rather than predefined features, provide a scalable pathway toward precision-driven intervention strategies for aging populations. - resting-state fMRI - foundation models - cognitive aging - intervention response - neuroplasticity Competing Interest Statement The authors have declared no competing interest. Footnotes ↵* Y. M. Liu Email: merik.liu{at}stanford.edu Competing Interest Statement: Authors claimed no conflict of interest. The revised manuscript has been extensively reworked in several important ways. First, we reframed the study. It is now a framework for discovering robust and generalizable latent neural targets of intervention response in cognitive aging, rather than primarily a model-evaluation paper. Second, we clarified the methodological logic. We explicitly positioned model fine-tuning as a clinically informed representation adaptation step that comes before downstream intervention-response prediction and latent pattern discovery. Third, we substantially revised and expanded the post hoc analyses. The new analyses better characterize latent brain-behavior structure and cross-study consistency. These now include multivariate decomposition of foundation model embeddings and their relationship to episodic memory change across two independent intervention trials. Fourth, we formalized a three-condition framework. This includes rich spatiotemporal representation, behavioral relevance, and cross-study robustness for clinically meaningful latent target discovery. This framework now guides the revised title, abstract, introduction, discussion, and figures.