FAST-Brain: A Flow-Aligned Spatio-Temporal Surrogate Brain Model
Abstract
Modeling resting-state functional magnetic resonance imaging (rs-fMRI) data is crucial for understanding brain-wide neural activity. However, traditional methods struggle to capture complex temporal dynamics over long horizons, to account for the brain's anatomical spatial structure, and to model high-dimensional ambient signals that lie on a low-dimensional intrinsic subspace. We propose FAST-Brain, a unified flow-aligned spatio-temporal surrogate brain model that addresses all three challenges. At its core is a flow-aligned generative framework that directly predicts the clean blood-oxygen-level-dependent (BOLD) signal, paired with a graph convolutional network that captures spatial structural constraints and a Transformer that models long-range temporal dependencies. Theoretically, we show that under a low-dimensional subspace assumption, the approximation error of our model scales with the intrinsic dimension rather than the ambient dimension, which justifies our direct modeling of the BOLD signal. Extensive experiments on synthetic and Human Connectome Project datasets demonstrate that FAST-Brain achieves state-of-the-art performance in recovering functional connectivity, effective connectivity, and the implicit low-dimensional signal subspace.
Related
- Routing to the seizure in postoperative human glioma cortex: spatio-temporal uncoupling between slow and fast extracellular neuronal activities
- Diffusion-enhanced spatio-temporal models for estimating spatial expansion
- 4D Spatio-temporal Capture of Leaf Morphogenesis Reveals Synchronisation between Growth and Motion
- Spatio-temporal visualisation and manipulation of endogenous Yap1 in medaka
- Cortical bursts provide spatio-temporal frames for the large-scale phase synchrony emerging across the birth transition
BibTeX
@misc{liu2026fastbrain,
title = {{FAST-Brain: A Flow-Aligned Spatio-Temporal Surrogate Brain Model}},
author = {Shucheng Liu and Changchun Shi and Kai Zhang and Hongtu Zhu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.34354},
eprint = {2609.34354},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.34354},
}
RIS
TY - UNPB
TI - FAST-Brain: A Flow-Aligned Spatio-Temporal Surrogate Brain Model
AU - Shucheng Liu
AU - Changchun Shi
AU - Kai Zhang
AU - Hongtu Zhu
PY - 2026
DA - 2026/09/28
PB - arXiv
DO - 10.48550/arxiv.2609.34354
UR - https://arxiv.org/abs/2609.34354
AB - Modeling resting-state functional magnetic resonance imaging (rs-fMRI) data is crucial for understanding brain-wide neural activity. However, traditional methods struggle to capture complex temporal dynamics over long horizons, to account for the brain's anatomical spatial structure, and to model high-dimensional ambient signals that lie on a low-dimensional intrinsic subspace. We propose FAST-Brain, a unified flow-aligned spatio-temporal surrogate brain model that addresses all three challenges. At its core is a flow-aligned generative framework that directly predicts the clean blood-oxygen-level-dependent (BOLD) signal, paired with a graph convolutional network that captures spatial structural constraints and a Transformer that models long-range temporal dependencies. Theoretically, we show that under a low-dimensional subspace assumption, the approximation error of our model scales with the intrinsic dimension rather than the ambient dimension, which justifies our direct modeling of the BOLD signal. Extensive experiments on synthetic and Human Connectome Project datasets demonstrate that FAST-Brain achieves state-of-the-art performance in recovering functional connectivity, effective connectivity, and the implicit low-dimensional signal subspace.
ER -