New preprint (bioRxiv 2026)
I am thrilled to announce the preprint “The practical impact of numerical variability on structural MRI measures of Parkinson’s disease“ is available on bioRxiv.
Description
This preprint quantifies numerical variability in structural MRI measures of Parkinson’s disease by instrumenting FreeSurfer to simulate numerical differences across computational environments. It introduces the Numerical-Population Variability Ratio (NPVR) to assess how numerical noise can influence reported group differences and clinical associations.
Authors
- Yohan Michel Benoit Chatelain
- Andrzej Sokolowski
- Madeleine Sharp
- Jean-Baptiste Poline
- Tristan Glatard