Skip to main content

Results

Overview​

The recorder captures a software run so it can be inspected, replayed and exported. Open an experiment, run at least one step, then choose 3 Review. Export and save controls remain disabled until there are recorded steps.

Recorded Data​

The recorder captures information including:

  • Simulation step
  • Simulation time
  • Demonstration agent position
  • Current region
  • Active stimuli
  • Treatment events
  • User interventions

Recording these values provides a chronological representation of the experiment.

Treatment Events​

Treatment events record information about treatment administration during an experiment.

The experiment can distinguish between:

  • No treatment
  • Placebo
  • Medicine

Treatment information includes the configured dose and administration timing where applicable.

Export Formats​

Results can be exported in two formats.

JSON​

JSON provides structured experiment data.

It is suitable for:

  • Programmatic processing
  • Reproducing experiment configurations
  • Further software integration
  • Python-based analysis

CSV​

CSV provides tabular experiment data.

It is suitable for:

  • Spreadsheet applications
  • Statistical analysis
  • Data visualisation
  • Import into Python and other analysis tools

Review and save a run​

  1. Read the run summary and event counts in Review. This is a quick inspection, not a statistical conclusion.
  2. Choose Read JSON in browser to see formatted data without downloading. Use Download JSON for the structured record or Download CSV for rows suitable for a spreadsheet.
  3. Choose Save this run to add the recording to Saved runs. A member needs write access; an observer cannot save a replay.
  4. Select a saved run to load its replay. The arena displays a Saved replay label, timeline, Play/Pause, Restart and Exit replay controls. Move the timeline to inspect a particular frame. Exit replay before starting a new live run.

If a download does not appear, check the browser's download permissions and any error beside the Review controls. Do not assume the file was saved merely because you clicked the button. If a saved run is missing, verify the account, project and experiment you are viewing.

Determinism Testing​

The project includes automated tests for reproducibility.

A deterministic test runs the simulation using a defined configuration and random seed and verifies that the resulting behaviour is consistent.

This helps ensure that a software trajectory is not unintentionally affected by display speed. It does not establish scientific validity of the demonstration agent or a treatment effect.

Using Results for Further Analysis​

Exported results can be used by external tools for additional analysis.

An analysis workflow is:

RODENT
      ↓
JSON / CSV
      ↓
Python
      ↓
Statistical Analysis
      ↓
Machine Learning / PyTorch

This is a pathway for future behavioural analysis and machine learning work. PyTorch training is not part of the current application. Researchers must validate the data and model assumptions before treating exported trajectories as biological evidence.

Result Verification​

Before using exported results, users should verify:

  • The correct experiment configuration was used.
  • The intended random seed was used.
  • The expected stimuli were activated.
  • Treatment events occurred at the expected times.
  • Recording completed successfully.
  • The exported file contains the expected data.

AI Attribution: This document was drafted and edited with AI assistance, including ChatGPT and OpenAI Codex. The team reviewed it against the implemented project.