arXiv:2605.26423v1 Announce Type: new Abstract: Task-based fMRI provides a direct readout of task-evoked neural dynamics, but it is expensive and difficult to acquire at scale, motivating rest-to-task synthesis from widely available resting-state fMRI (rsfMRI). We propose FM-fMRI, an event-conditioned flow-matching model that learns a continuous-time conditional vector field to generate task ROI time series from a subject's rsfMRI and the task event information. The formulation enables fast ODE-based sampling and flexible conditioning over heterogeneous event schedules. Rather than optimizing
Advances in AI, specifically flow-matching models, are enabling sophisticated synthesis tasks in medical imaging, coinciding with increased computational power and demand for medical data efficiency.
This development could significantly reduce the cost and logistical challenges of acquiring task-based fMRI data, accelerating neuroscience research and clinical applications by making detailed neural dynamics more accessible.
The ability to synthesize task-based fMRI data from resting-state fMRI changes how researchers can study brain function, potentially democratizing access to complex neural activity insights.
- · Neuroscience researchers
- · fMRI manufacturers (increased demand for base rsfMRI)
- · AI model developers
- · Pharmaceutical R&D
- · Traditional task-based fMRI acquisition services (potentially reduced demand)
- · Studies relying solely on expensive, large-scale task fMRI cohorts
More widespread and cost-effective analysis of task-evoked neural dynamics becomes possible.
Accelerated discovery of biomarkers for neurological and psychiatric conditions due to expanded datasets.
The development of highly personalized medical interventions based on synthesized individual brain responses to various tasks.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG