Resources are organized by version, starting with v0.0.3. Scripts to generate these resources are available at https://github.com/neurostuff/nimare-resources.
- 0.0.3:
- `neurosynth_with_cogat.pkl.gz`: NiMARE Dataset containing Neurosynth database with the following features:
- Neurosynth tf-idf features
- abstracts
- Cognitive Atlas term counts
- expanded Cognitive Atlas terms with simple hierarchical expansion.
- `neurosynth_gclda.pkl.gz`: GCLDA model trained on Neurosynth database with count version of Neurosynth features, with the following parameters:
- 200 topics
- symmetric regions
- 50000 iterations
- other parameters at defaults