Latest data news

AI & data science

DEGAS uses brain networks to classify Parkinson’s subtypes

A study reports that DEGAS classified early Parkinson’s patients by suspected disease course using baseline functional MRI connectivity.

What happened

The npj Parkinson’s Disease study analysed baseline scans from 336 early-stage patients. Its graph-based classifiers achieved balanced accuracy of 0.91–0.93, compared with 0.79 for adjacency-only models.

Why it matters

For clinical data scientists and neuroimaging analysts, network features may help distinguish patients at risk of faster progression. Those classified as having suspected REM sleep behaviour disorder had worse outcomes over three years.

What to watch

Teams could check whether the network signatures replicate in an independent patient cohort before considering clinical use.