Resources & Collaboration

Access the datasets, documentation, and engagement channels that power our progressive AI program. We emphasize responsible data stewardship and transparent sharing.

Core datasets

PPMI (Parkinson’s Progression Markers Initiative)

Primary longitudinal cohort with clinical, imaging, biospecimen, and genetic data. Requires data use agreement via Michael J. Fox Foundation.

Access resource

synapse.org Parkinson’s repositories

Wearable, smartphone, handwriting, and voice datasets curated for open challenges and benchmarking.

Access resource

Mindboggle-101

Anatomically labeled T1-weighted MRIs used for SegFormer pretraining and validation.

Access resource

Lab documentation

Explore protocol drafts, segmentation benchmarks, and reproducible notebooks on our background page and HackMD workspace. Highlights include:

Open-source roadmap

SegFormer deployment scripts

Containerized pipelines for MRI preprocessing, segmentation, DaT-SPECT alignment, and SBR extraction. Repository release planned December 2025.

Progressive agent notebooks

Reference implementations of stochastic Hamiltonian policy optimization and uncertainty calibration on synthetic cohorts, packaged with the upcoming technical note.

ActionIntel & Free Motion demos

Interactive Dash applications highlighting sensor-clinical correlations, counterfactual therapy simulations, and visualization of asymmetric gait patterns.

Engage with us

Connect with Professor Bajaj

Share potential collaborations or student opportunities directly with our principal investigator.

Faculty profile

Oden Institute contact portal

For media inquiries, philanthropy, or institutional partnerships, please contact the Oden Institute communications team.

Open contact page

Texas Advanced Computing Center

External partners can apply for compute allocations that align with our GPU-intensive workloads.

Request resources

Responsible data stewardship

All collaborations comply with HIPAA, GDPR, and sponsor-specific requirements. We utilize secure UT Austin data enclaves, audited access logs, and federated learning when raw data sharing is constrained.