Omni-fMRI: A Universal Atlas-Free fMRI Foundation Model

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Mo, Ye, Wenhao, Xia, Junfeng, Zhang, Junxiang, Pan, Xuanye, Xu, Minghao, Deng, Haotian, Wen, Hongkai, Liu, Quanying
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910006235889664
author Wang, Mo
Ye, Wenhao
Xia, Junfeng
Zhang, Junxiang
Pan, Xuanye
Xu, Minghao
Deng, Haotian
Wen, Hongkai
Liu, Quanying
author_facet Wang, Mo
Ye, Wenhao
Xia, Junfeng
Zhang, Junxiang
Pan, Xuanye
Xu, Minghao
Deng, Haotian
Wen, Hongkai
Liu, Quanying
contents Self-supervised fMRI foundation models have shown promising transfer performance, yet most rely on predefined region-level parcellations that discard fine-grained voxel information and introduce atlas-dependent biases. We propose Omni-fMRI, an atlas-free foundation model that operates directly on voxel-level signals. To enable scalable pretraining on 49,497 fMRI sessions across nine datasets, Omni-fMRI introduces a dynamic patching mechanism that substantially reduces computational cost while preserving informative spatial structure. To support reproducibility and fair comparison, we establish a comprehensive benchmark suite spanning 11 datasets and a diverse set of resting-state and task-based fMRI tasks. Experimental results demonstrate that Omni-fMRI consistently outperforms existing foundation models, providing a scalable and reproducible framework for atlas-free brain representation learning. Code and logs are available.
format Preprint
id arxiv_https___arxiv_org_abs_2601_23090
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Omni-fMRI: A Universal Atlas-Free fMRI Foundation Model
Wang, Mo
Ye, Wenhao
Xia, Junfeng
Zhang, Junxiang
Pan, Xuanye
Xu, Minghao
Deng, Haotian
Wen, Hongkai
Liu, Quanying
Computational Engineering, Finance, and Science
Quantitative Methods
Self-supervised fMRI foundation models have shown promising transfer performance, yet most rely on predefined region-level parcellations that discard fine-grained voxel information and introduce atlas-dependent biases. We propose Omni-fMRI, an atlas-free foundation model that operates directly on voxel-level signals. To enable scalable pretraining on 49,497 fMRI sessions across nine datasets, Omni-fMRI introduces a dynamic patching mechanism that substantially reduces computational cost while preserving informative spatial structure. To support reproducibility and fair comparison, we establish a comprehensive benchmark suite spanning 11 datasets and a diverse set of resting-state and task-based fMRI tasks. Experimental results demonstrate that Omni-fMRI consistently outperforms existing foundation models, providing a scalable and reproducible framework for atlas-free brain representation learning. Code and logs are available.
title Omni-fMRI: A Universal Atlas-Free fMRI Foundation Model
topic Computational Engineering, Finance, and Science
Quantitative Methods
url https://arxiv.org/abs/2601.23090