Clinicians and researchers use diffusion MRI to study how brain tissue is organized and how disease affects it by tracking the movement of water molecules through the brain.

Traditionally, these scans are interpreted under the assumption that each small 3D block of the image (called a voxel) is uniform—that is, that tissue properties affecting the scan signal are the same throughout that unit. In reality, brain tissue is more complex.

Iman Aganj, PhD

A new study led by Iman Aganj, PhD, of the Athinoula A. Martinos Center for Biomedical Imaging at Mass General Brigham, examined whether this assumption causes important information to be missed.

The researchers hypothesized that as water molecules diffuse through tissue, they encounter subtle changes in their local surroundings that slightly influence the diffusion MRI signal.

Using a new mathematical framework, they used these small signal variations to infer how tissue properties change within a voxel, rather than relying only on contrasts between neighboring voxels.

The team tested the method on multiple datasets, including large collections of existing human brain diffusion MRI scans as well as specially designed scans with very long diffusion times that amplify the effect (which they share with the public). Across all datasets, the approach recovered information inside voxels that standard models typically miss.

If used to complement standard diffusion MRI analysis, this new math could improve understanding of tissue microstructure, aid the discovery of new biomarkers, and increase the effective resolution in diffusion‑based brain imaging.

Other contributors to this research are Thorsten Feiweier, John Kirsch, Bruce Fischl and André van der Kouwe.

Published in Magnetic Resonance in Medicine on January 23, 2026 | Read the paper.

Summary reviewed by: Iman Aganj, PhD, lead author