#neuroimaging

#neuroimaging

According to University of Michigan neuroscientists, not only can their AI vision language model diagnose neurological disorders from MRI scans with high performance accuracy, but it also has foundation model capabilities, making it a flexible, general-purpose solution that can be tailored for a wide variety of medical imaging. "These results demonstrate that Prima has foundation model properties, and reported performance will continue to improve with additional health system training data and larger compute budgets," wrote the study's authors in the preprint.

Before treatment began, participants underwent neuroimaging. Instead of relying on a single modality, the researchers fused structural connectivity (how regions are physically wired) with functional connectivity (how regions co-activate at rest). The goal was not to throw every possible feature at a black box, but to learn a constrained pattern-what the authors call structure-function "covariation"-that carries the most predictive signal for outcome. In other words, the model tries to find the smallest set of connections that meaningfully forecasts symptom change.