Researchers adapt Vision Transformers for fMRI analysis using flat maps

Researchers have developed a new family of models called CortexMAE, which adapt Vision Transformers for analyzing functional MRI data by projecting 3D volumes into 2D flat maps. This approach, tested on over 2,000 hours of fMRI data, shows power-law scaling characteristics and has led to the creation of the first open evaluation suite for fMRI foundation models, named Brainmarks. While the models struggle with subject-level trait prediction, they demonstrate significant improvements in cognitive state decoding compared to previous methods. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Introduces new methods for analyzing medical imaging data, potentially improving diagnostic capabilities and understanding of brain function.

RANK_REASON The cluster contains a new research paper detailing a novel model architecture and evaluation suite for fMRI data analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• CortexMAE
• Vision Transformer
• fMRI
• Brainmarks
• Connor Lane
• arXiv

• paper
• model release
• infra