Quick start¶
Launch the GUI¶
conda activate cicada_env
python -m cicada_gui
No data at hand? Select group > Demo files loads the demo files installed with CICADA, or open files from DANDI with File > Open new dataset from DANDI... (see Data sources).
Minimal workflow¶
cicada-gui main window with example NWB loaded and sorted by 'Subject ID'
Minimal use workflow¶
1. Load sessions¶
Click Open new dataset using the File menu and select the directory containing
your .nwb files. All valid NWB files are listed in the sessions panel.
2. Select sessions¶
Tick the sessions you want to include in the next analysis run.
Use Group by from the View menu to sort the list by metadata field (age, genotype, sex, subject ID, session type, etc.) — useful for quickly selecting a subset of sessions.
3. Save / load a group¶
Create group in the File menu writes a sessions_<name>.yaml file to your configured result
directory (see Configuration in the GUI guide). This file can be reloaded later (Load group) or used directly in
the headless workflow.
Tip: the Load group option can replace Open folder — it will load only the files listed in the selected group.
4. Pick an analysis¶
The analysis tree shows all available plugins, organized by family. Click a family header to expand it.
Analyses that are incompatible with the currently checked sessions are greyed out
(the cicada-analysis - check_data() guard reported a problem — hover to see why).
5. Open the parameter form¶
Double-click an analysis to open its parameter window. The form is organized into collapsible sections (families).
cicada-gui analysis window for the cicada-analysis 'Display cell maps' plugin.
Session-specific parameters (ROI response series, epoch names, etc.) are pre-filled from the checked sessions.
6. Save / load parameters¶
- Save parameters — writes a
params_<ClassName>.yamlto disk. Compatible with the headless runner: use it as theparamsargument ofrun_from_config. - Load parameters — loads a previously saved
params_<ClassName>.yamlorcombined_*.yaml. Session-specific parameters (ROI response series, etc.) are skipped automatically when loading across different session groups.
7. Run¶
Fill all mandatory parameters (highlighted if missing), then click RUN ANALYSIS.
After the run:
- Results are saved to the results_path parameter you set.
- A combined_<ClassName>_<timestamp>.yaml file is written alongside the results —
it records the exact sessions and parameters used, enabling headless replay.
Same analysis, headless¶
The same analysis plugin runs identically in both modes:
# GUI mode
cicada_gui
# Headless mode — same analysis, same parameters, no GUI required
python templates/run_from_config_template.py
See cicada-analysis for the full headless workflow.
More details on every window: GUI guide.