Robust automated preclinical fMRI preprocessing via a multi-stage dilated convolutional Swin Transformer affine registration

Sima Soltanpour, Md. Taufiq Nasseef, Rachel Utama, Arnold Chang, Dan Madularu, Praveen Kulkarni, Craig F. Ferris, Chris Joslin · 2025

Read Robust automated preclinical fMRI preprocessing via a multi-stage dilated convolutional Swin Transformer affine registration by Sima Soltanpour, Md. Taufiq Nasseef, Rachel Utama, Arnold Chang, Dan Madularu, Praveen Kulkarni, Craig F. Ferris, Chris Joslin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-12742213.txt.

Description

Sima Soltanpour and co-authors assemble, in a 2025 paper, a fully automated pipeline that prepares animal brain scans for analysis, ending the hand-tuning that slows preclinical imaging. Functional scans of small animals arrive noisy, low in resolution, and geometrically unlike the atlases they must be matched to. Their pipeline chains three learned modules: a generative network that denoises, a transformer that strips the skull, and a new registration network that warps each brain onto the atlas.

The registration module, a multi-stage Swin transformer widened with dilated convolutions, reads misalignment at both the local and global scale, which is where earlier methods slipped. Across several preclinical data sets it aligned brains more accurately, by the standard overlap score, than the leading alternatives. The authors frame the pipeline as robustness work: each stage answers a failure mode the data actually present.

Catalog record

CREATORSima Soltanpour, Md. Taufiq Nasseef, Rachel Utama, Arnold Chang, Dan Madularu, Praveen Kulkarni, Craig F. Ferris, Chris Joslin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC12742213

Browse the full library · About the Tower of Babel