Adrian Dalca News Today

Tyche is a machine-learning framework that can generate plausible answers when asked to identify potential disease in medical images. By capturing the ambiguity in images, the technique could prevent clinicians from missing crucial information that could inform diagnoses.

Compared to other imaging modalities like X-rays or CT scans, MRI scans provide high-quality soft tissue contrast. Unfortunately, MRI is highly sensitive to motion, with even the smallest of movements resulting in image artifacts. These artifacts put

MIT researchers develop an algorithm combining deep learning and physics for motion correction in MRI scans of the brain. Experts say this could help children, elderly, and people with psychiatric or neurological disorders who may have difficulty staying still during an MRI scanning session.