How genevievewager/hippo works
HIPPO is a hippocampal BCI simulation, decoder-search, and deployment testbed. Its main workflow simulates behavior and neural populations, models acquisition degradation and sorting, constructs causal spike observations, compares static and dynamic representations and decoders, selects deployable models, then evaluates replay and closed-loop policies. Users enter through the CLI scripts or Streamlit UI; visualization inspects saved results rather than retraining. The graph follows the documented workflow and sampled pipeline-state/UI code; several scientific stage internals were not sampled, so their exact call wiring is not expanded.
Main entry points: dataset.py Experiment dataset · run_simulation.py CLI workflows · app.py Streamlit app · extractor.py Neural features
Components 21
User entry
- CLI workflows
run_simulation.py - Streamlit app
ui/app.py
Simulation
- Behavior simulation
hippo_sim/behavior.py - Hippocampal populations
hippo_sim/hippocampal_populations.py - Probe anatomy
hippo_sim/anatomy.py - Acquisition and sorting
hippo_sim/pipeline.py - Experiment dataset
hippo/dataset.py
Neural decoding
- Causal observations
realtime/observation.py - Neural features
realtime/neural_features/extractor.py - Static manifolds
realtime/manifolds/registry.py - Dynamic latents
realtime/dynamic_latents/registry.py - Decoder search
realtime/decoder_comparison.py
Deployment
- Deployment selection
realtime/deployment_selection.py - Lab registry
realtime/lab_deployable.py - Realtime replay
realtime/live_decoder.py - Closed-loop policy
realtime/closed_loop_controller.py - Live bundle
realtime/deployment_bundle.py
State and inspection
- Pipeline provenance
realtime/pipeline_graph.py - Saved-result figures
visualization/load_outputs.py
Other components
- Experimenter
- Open Ephys
Connections 24
- Experimenter → CLI workflows: runs workflow
- Experimenter → Streamlit app: uses pages
- CLI workflows → Behavior simulation: starts simulation
- Behavior simulation → Hippocampal populations: drives activity
- Hippocampal populations → Probe anatomy: uses anatomy
- CLI workflows → Acquisition and sorting: runs pipeline
- Acquisition and sorting → Experiment dataset: writes dataset
- Experiment dataset → Causal observations: supplies spikes
- Causal observations → Neural features: constructs inputs
- Neural features → Static manifolds: fits representations
- Neural features → Dynamic latents: fits latent states
- Static manifolds → Decoder search: provides embeddings
- Dynamic latents → Decoder search: provides states
- Decoder search → Deployment selection: scores candidates
- Deployment selection → Lab registry: exports winners
- Lab registry → Realtime replay: loads models
- Realtime replay → Closed-loop policy: supplies predictions
- Deployment selection → Live bundle: packs deployment
- Live bundle → Open Ephys: targets integration
- Streamlit app → Experiment dataset: selects dataset
- Streamlit app → Pipeline provenance: reads status
- Pipeline provenance → Experiment dataset: tracks stages
- CLI workflows → Saved-result figures: visualizes outputs
- Experiment dataset → Saved-result figures: loads saved results