Nested RF and Direction Selectivity Protocols

Generates, runs and analyses two-stage receptive-field and direction-selectivity protocols for patch-electrophysiology recordings of Drosophila T4/T5 neurons on the G4 LED arena.

Version README v1 · Status active · Last verified not yet verified

What this does

• Protocol 1 — presents pre-made flash grids to find the rough receptive-field location and preferred contrast of the recorded cell, giving a peak_frame.
• Protocol 2 — generates a bespoke protocol centred on that peak_frame: high-resolution RF mapping with small flashing squares, plus direction selectivity from moving bars.
• Analysis — extracts direction-selectivity metrics and 2-D Gaussian receptive-field fits, and produces the manuscript figures.

You will need: a G4 LED arena, connected and calibrated · G4_Display_Tools on the MATLAB path · MATLAB (Windows, for the rig) · recorded data — none is included in this repo.

Quick start

Running an experiment needs the rig. To work with data you already have, start at Workflow step 3.

1. Get the code

   git clone https://github.com/leburnett/nested_RF_stimulus.git

2. Add the source tree to the MATLAB path

   addpath(genpath('src'))

3. Generate and run protocol 2

   generate_protocol2()

   Prompts for the peak_frame from protocol 1, the arena side, and the fly's age and strain; writes a timestamped experiment folder (Patterns/, Functions/, currentExp.mat) under <matlabroot>\G4_Protocols\nested_RF_protocol2\; runs it on the arena; prints the inverse peak frame for the opposite-contrast experiment.

Repository map

Directory	Contents
src/protocol_generation/	Builds and runs protocol 2 — details
src/stimulus_generation/	Flash, bar and position-function generation for both protocols
src/analysis/	process_protocol2.m and the bar/flash/RF analyses
src/preprocessing/	Batch result generation and dashboard image export
src/dashboard/	Dash app for browsing per-cell results (pixi environment)
scripts/	Manuscript figures — details
protocols/	Pre-made protocol 1 .g4p files, by arena side and background level
results/	Pre-made patterns, position functions and generated outputs
docs/	Protocol design and history

Workflow

The Quarto guide is the canonical step-by-step version. In brief:

1. Present protocol 1 — run a pre-made .g4p from protocols/ (the folder matching the arena side and background level) via G4_experiment_conductor; read the peak_frame off the response plots.
2. Generate and run protocol 2 — generate_protocol2(), as above.
3. Process the data — from inside the experiment folder:

   cd('<path to>/nested_RF_protocol2/data/<timestamp>')
   process_protocol2()

   Writes bar_results/ and flash_results/, plus figures. Inputs and outputs.
4. Browse and plot — cd src/dashboard && pixi run dashboard (http://localhost:8051); manuscript figures from scripts/.

Key data types

Output	Description	Units
bar_results_*.mat	Responses to 16 bar directions at 3 speeds, DSI, preferred direction	mV; deg
rf_results_*.mat	2-D Gaussian RF fit, excitatory and inhibitory components	mV; arena px
Bar speeds	Sweeps presented at three speeds	28, 56, 168 deg/s

Stimulus grids and timing: src/analysis/README.md.

Conventions & gotchas

• Bar stimuli are not regenerated. Everything else is built de novo by generate_protocol2(), but the bar sweep and bar flash stimuli use pre-made patterns in results/patterns/protocol2 — change their parameters by editing those patterns directly.
• process_protocol2() must be run from inside the experiment directory, not the repo.
• Protocol folders are named yyyy_MM_dd_HH_mm at generation time.
• Several analysis scripts still contain hardcoded data paths; edit before running elsewhere.

Related

• Full instructions and troubleshooting — the canonical step-by-step guide.
• G4_Display_Tools — arena control.

Contact

Laura Burnett, Reiser Lab, HHMI Janelia Research Campus.