Stanford Longitudinally Accelerated MRI Dataset (SLAM)

We created the Stanford Longitudinally Accelerated MRI (SLAM) dataset to support our work on diffusion-based longitudinal MRI reconstruction, where prior scans are integrated to accelerate and improve follow-up imaging. SLAM currently contains 232 brain MRI scans from 53 subjects, including raw multi-coil k-space, corresponding complex-valued reconstructions, and paired prior DICOM scans across a variety of neuroimaging protocols, contrasts, intervals, and pathologies. We are actively collecting additional data, planning to expand this dataset for future releases. By releasing SLAM publicly, we aim to provide the research community with a resource tailored for benchmarking and advancing methods in longitudinal MRI reconstruction, a critical yet underexplored area of clinical imaging.

Note: Bulk download of this dataset may be disabled, so please alternatively use Globus at this link for more reliable download: https://app.globus.org/file-manager?origin_id=0e3efb9a-ad98-425f-8b84-4247281c91f6&origin_path=%2Fsdr_access%2Frq%2F296%2Frb%2F2765%2F&two_pane=true.
This will require setting up an account with Globus and installation of the free Globus Connect Personal software for bulk transfer.