The Brain Abnormality Workshop
Advancing computational neuroimaging and computational neuropathology methods that remain reliable in the presence of tumors, traumatic brain injury, infection, diffuse pathology, and imaging artifacts.
Why BrainWorks?
BrainWorks aims to advance computational neuroimaging and computational neuropathology methods that operate effectively in the presence of brain abnormalities. Many machine learning methods in neuroimaging were developed and validated using neurologically healthy cohorts, which can lead to reduced performance when anatomy, tissue integrity, or signal characteristics depart from normative assumptions.
BrainWorks creates a forum for pathology-aware, generalisable image computing methods that explicitly address tumors, acquired brain injuries, enlarged perivascular spaces, infection, diffuse pathology, and imaging artifacts. The workshop emphasizes both methodological innovation and careful analysis of failure modes in clinically realistic data.
Primary aims
Call for papers
We invite submissions on all aspects of medical image computation and computer assisted interventions involving brains with abnormalities, from focal lesions to diffuse pathology and artifacts.
Submission instructions
We invite submissions of original research papers. All submissions will undergo a double blind peer review process. Authors are encouraged to submit early, as decisions may be made on a rolling basis prior to the final submission deadline.
Papers must be fully anonymized before submission. A paper that is non-compliant with the anonymization or other paper submission rules (e.g., paper length) will be desk rejected.
Submissions that have been previously published, or are currently under review at other venues are not permitted. However, authors participating in MICCAI 2026 Challenges that do not produce their own proceedings (e.g., SELMA3D, TopBrain, DoseRAD2026, FOMO26) are welcome to submit their work to BrainWorks. If BrainWorks submission preceeds the release of challenge results, authors may update the final leaderboard standing or ranking in the camera-ready version to reflect the official challenge outcome, provided no other substantive changes to the results are made. Authors submitting work that is also submitted to a MICCAI 2026 Challenge are required to declare this at the time of submission.
An open-access version of all accepted papers from the MICCAI 2026 Satellite Event BrainWorks will be made available on the MICCAI Society website no earlier than one week before the first day of the conference. Authors intending to file patents are responsible for ensuring that all necessary filings are completed prior to this public release.
Program coming soon
Whole day thematic event encompassing AIMS-TBI and BraTS 2026.
The AIMS-TBI 2026 challenge aims to develop a lesion segmentation tool capable of handling immense heterogeneity among lesion types. This year AIMS-TBI includes multi-modal data for training, with segmentation performance being assessed on T1w images only.
The BraTS 2026 Cluster of challenges includes three segmentation challenges (Metastasis, Pediatrics and Generalizability), the BraTS Pathology and BraTS Inpainting challenges.
Organizing team
BrainWorks is organized by a cross-institutional team spanning early-, mid-, and senior-career researchers with expertise in neuroimaging, computer vision, machine learning, and clinical applications.
Organizing committee
evelyn.deutscher@mcri.edu.au (Primary contact)
Advisory board
Program committee
FAQ
Researchers and clinicians working on computational neuroimaging, neuropathology, biomarkers, segmentation, registration, classification, harmonization, multimodal fusion, robustness, and related topics in abnormal brain data.
Yes. BrainWorks explicitly welcomes studies that analyze limitations, failure cases, and gaps between benchmarks and real-world deployment.
Yes. The workshop plans proceedings in the MICCAI Satellite Events Joint Springer LNCS proceedings.
BrainWorks is the broader scientific forum spanning multiple brain abnormalities and methodological questions, complementing task-specific challenge benchmarking efforts.