Brain Foundation Models
Standard fine-tuning of foundation models was less hub-sensitive than BRAINMAP. Foundation models showed stronger generic explanation robustness than BRAINMAP under RE-CONFIRM sanity checks. The study evaluates whether brain foundation mod…
1 sources - 4 claims
Standard fine-tuning of foundation models was less hub-sensitive than BRAINMAP. Foundation models showed stronger generic explanation robustness than BRAINMAP under RE-CONFIRM sanity checks. The study evaluates whether brain foundation models produce robust biomarkers from dynamic functional connectivity despite limited disease-state representation in pretraining. Brain foundation models are pre-trained on large brain-imaging datasets and can be adapted to downstream brain-analysis tasks.