When adding a new project to the listing, please copy and paste the template below.
Project url(s): https://github.com/galilej25/NBPy/
Contributors: Ljubomir Radakovic, Waleed Aamer, Zeus Gracia-Tabuenca.
Description of project: Network-Based Statistics framework in Python. NBpy allows to extract by statistical inference clusters of connections based on family-wise error permutation test.
How to get involved: if you are interested in network science, linear mixed-effects, and/or computing efficiency applied to large-scale networks with Python... please feel welcome to this project :)
Project url: https://github.com/neurohackademy/feature-encoding
Project Contributors (alphabetical): Alberto Mario Ceballos Arroyo, Tomas D'Amelio, Heejung Jung, Adriana Mendez Leal, Shawn Rhoads, Shelby Wilcox, Tiankang Xie
Description of project: Our project goals include (1) Using Facebook's SlowFast video classification model to extract features (e.g., kinetics) from video stimuli; (2) Building a feature encoding model using LASSO regression to examine which brain regions encode information about the observed features
- Project url(s): https://github.com/open-dicom/dicom_parser
- Contributors: Lily Cheng, Michał Szczepanik, Zvi Baratz
- Description of project:
- Documentation and Community
- Created Open DICOM organization on GitHub.
- Reorganized to be more accessible and easy to navigate.
- Created a “resources” page with many useful learning materials and other tools.
- Organized the differences in return types between pydicom and dicom_parser.
- Multi-frame Support
- Create the MultiFrame and FunctionalGroups classes.
- Complete coverage except for a specific Phillips-specific case for which we don’t yet have a sample file.
- Long discussion regarding test files management. WIP creating a unified testing resources.
- Final presentation
- Documentation and Community
- How to get involved: Check out Open DICOM and the DICOM discourse channel.
Project url(s): https://github.com/yukaizou2015/connectivityml
Contributors: Priya Kalra, Tal Pal Attia, Debbie Burdinski, Pradyumna Lanka, Yukai Zou
Description of project: This project utilized resting-state functional connectivity data from the Human Connectome Project to explore whether individual differences in networks predict individual differences in cognition and demographics, and whether particular edges are more informative. Machine learning algorithms, including random forest classifiers and support vector machine were employed on edge weighting features derived from independent component analysis, and the results from random forest classifiers were interpreted using SHAP values. Lastly, a framework was laid out for optimizing the machine learning algorithmn used to predict differences in cognition or demographics.
How to get involved: We are very keen on implementing new modeling approaches as well as getting our analyses reproducible and explainable. If you feel you can contribute, please do not hesitate to be in touch with us.