How gel1has3/neurobench_studio works
NeuroBench Studio is a Flask-based EEG research platform with a desktop/web launcher, dataset and model explorers, a conversational assistant, and a visual pipeline workflow. Users browse or select EEG data, configure and run analysis, then inspect experiment results and visualizations. The graph preserves these documented product branches; source excerpts support the launcher and several backend calls, while some dashboard integrations are documented only in the README, so their connections are marked dashed.
Main entry points: flask_app.py Flask dashboard · pipeline_executor.py Pipeline executor · base.html Browser interface · braindecode_explorer.html Model discovery
Components 18
Platform entry
- Desktop launcher
main.py - Flask dashboard
src/dashboard/flask_app.py - Browser interface
src/dashboard/templates/base.html
Exploration tools
- EEG assistant
src/dashboard/conversational_eval.py - Dataset explorer
src/dashboard/flask_app.py - Local EEG data
data/raw - Model discovery
src/dashboard/templates/braindecode_explorer.html - Foundation models
src/models/foundation.py
Analysis workflow
- Pipeline builder
src/dashboard/templates/pipeline_builder.html - Pipeline executor
src/dashboard/pipeline_executor.py - Dataset loaders
src/preprocessing/dataset_loaders.py - Braindecode wrappers
src/models/braindecode_models.py
Results and insights
- Evaluation metrics
src/evaluation/metrics.py - Representation analysis
src/evaluation/geometry_analysis.py - Experiment results
results - Interactive charts
src/dashboard/static/js/charts.js
Other components
- Researcher
- EEGDash service
Connections 19
- Researcher → Browser interface: uses
README.md - Desktop launcher → Flask dashboard: starts
main.py - Browser interface → Flask dashboard: requests pages
- Flask dashboard → EEG assistant: imports
src/dashboard/flask_app.py - Flask dashboard → Dataset explorer: serves
README.md - Flask dashboard → Pipeline executor: imports
src/dashboard/flask_app.py - Dataset explorer → Local EEG data: scans
README.md - Dataset explorer → EEGDash service: streams datasets
README.md - Researcher → Model discovery: discovers models
README.md - Model discovery → Foundation models: documents models
- Researcher → Pipeline builder: configures
README.md - Pipeline builder → Pipeline executor: runs pipelines
README.md - Pipeline executor → Dataset loaders: loads datasets
src/dashboard/pipeline_executor.py - Pipeline executor → Braindecode wrappers: uses models
src/dashboard/pipeline_executor.py - Pipeline executor → Evaluation metrics: evaluates
- Pipeline executor → Experiment results: records results
- Flask dashboard → Experiment results: loads results
src/dashboard/flask_app.py - Flask dashboard → Representation analysis: loads analysis
src/dashboard/flask_app.py - Flask dashboard → Interactive charts: supplies visualizations