# A Scaling Study for fMRI Foundation Models

arXiv: https://arxiv.org/abs/2609.27232 · HTML: https://mlsanity.com/paper/2609.27232/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 87 references that are in the map.
- First posted on arXiv 23 Sep 2026.

## Builds on (most cited first, 17 of 17)

- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [Reproducible scaling laws for contrastive language-image learning](https://mlsanity.com/paper/2212.07143/) (2022)
- [Scaling Vision Transformers](https://mlsanity.com/paper/2106.04560/) (2021)
- [Scaling Data-Constrained Language Models](https://mlsanity.com/paper/2305.16264/) (2023)
- [LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics](https://mlsanity.com/paper/2511.08544/) (2025)
- [BrainMass: Advancing Brain Network Analysis for Diagnosis with Large-scale Self-Supervised Learning](https://mlsanity.com/paper/2403.01433/) (2024)
- [Brain Harmony: A Multimodal Foundation Model Unifying Morphology and Function into 1D Tokens](https://mlsanity.com/paper/2509.24693/) (2025)
- [Exploring the reproducibility of functional connectivity alterations in Parkinson's Disease](https://mlsanity.com/paper/1611.04794/) (2016)
- [Omni-fMRI: A Universal Atlas-Free fMRI Foundation Model](https://mlsanity.com/paper/2601.23090/) (2026)
- [Brain-DiT: A Universal Multi-state fMRI Foundation Model with Metadata-Conditioned Pretraining](https://mlsanity.com/paper/2604.12683/) (2026)
- [SLIM-Brain: A Data- and Training-Efficient Foundation Model for fMRI Data Analysis](https://mlsanity.com/paper/2512.21881/) (2025)
- [FlexiBrain: Resolution-Agnostic Voxel-Level Encoding for Native fMRI](https://mlsanity.com/paper/2606.11500/) (2026)
- [BrainWorld: A Structural-Prior-Conditioned Generative Model for Whole-Brain 4D fMRI Dynamics](https://mlsanity.com/paper/2606.17742/) (2026)
- [Brain-Semantoks: Learning Semantic Tokens of Brain Dynamics with a Self-Distilled Foundation Model](https://mlsanity.com/paper/2512.11582/) (2025)
- [A Brain Graph Foundation Model: Pre-Training and Prompt-Tuning across Broad Atlases and Disorders](https://mlsanity.com/paper/2506.02044/) (2025)
- [BrainTaskonomy: Learning How to Pretrain and What to Transfer in fMRI Foundation Models](https://mlsanity.com/paper/2609.10518/) (2026)

## Abstract (arXiv)

Scaling laws have guided large-model development in computer vision and natural language processing, but the relationships among data, model size, and compute remain unclear for functional magnetic resonance imaging (fMRI) foundation models. Here, we conduct a controlled empirical study using pretraining data from more than 200 source datasets and over 10,000 GPU-hours of experiments. Holding the pretraining framework and downstream protocol fixed, we vary pretraining data size, model size, and training duration. Downstream performance generally improves with compute, yet models using similar compute can perform substantially differently. Additional pretraining data bring larger gains at larger model sizes, suggesting that data and model size should be scaled together. At matched compute, increasing pretraining data benefits more tasks than increasing model size, although the pattern varies across tasks. We then use in-distribution (ID) downstream performance to select the combination of pretraining data size, model size, and training duration at two fixed compute budgets. The resulting models are locked before out-of-distribution (OOD) evaluation. They achieve the highest average performance across the evaluated OOD tasks among the compared fMRI foundation models while using less pretraining compute. Overall, our results show that compute alone does not characterize fMRI scaling: performance depends on how pretraining data, model size, and training duration are combined.