Repository of clinical brain scans with real-world variability

This is the "MGH dataset" used in the article: "SynthSR: a public AI tool to turn heterogeneous clinical brain scans into high-resolution T1-weighted images for 3D morphometry" Juan E. Iglesias, Benjamin Billot, Yaël Balbastre, Colin Magdamo, Steven E. Arnold, Sudeshna Das, Brian L. Edlow, Daniel C. Alexander, Polina Golland, and Bruce Fischl Science Advances, eadd3607(2023) * The file subjects.tar contains 15926 scans from 1384 subjects in compressed nifti format (*.nii.gz). * The spreadsheet demographics.csv contains the age and gender of the subjects. * The zip file SynthSR-main-Dec_8_2022.zip contains a snapshot of the source code from GitHub (https://github.com/BBillot/SynthSR) at the type of publication. * All other PHI has been removed with anonymization purposes. * The scans have been skull stripped with SynthSeg+ (please see: https://surfer.nmr.mgh.harvard.edu/fswiki/SynthSeg). * These are MRI data "in the wild", straight from the MGH picture archiving and communication system (PACS). We have included all scans that ran through SynthSeg, including plenty of scans that are unusable due to limited field of view, excessive noise, non-structural MRI sequence, etc. Every subject has a different number of scans with different MRI sequences, resolutions, orientations, etc.