Specialized Parallel Pathways for Adaptive Control of Visual Object Pursuit

This repository accompanies the manuscript:

Collie, M.F., Jin, C., Kellogg, E., Vanderbeck, Q.X., Hartman, A.K., Holtz, S.L., & Wilson, R.I.
Specialized parallel pathways for adaptive control of visual object pursuit
bioRxiv, 2025.

---

Overview

This project investigates the neural basis of adaptive visual pursuit in Drosophila melanogaster, focusing on two distinct AOTU pathways (AOTU019 and AOTU025) that enable context-sensitive steering behavior. Using connectomics, electrophysiology, behavior, and modeling, we show how feedback loops implement flexible gain control during pursuit.

---

Repository Contents

The repository is organized into the following directories:

1. Connectomics_Code/

Code for analyzing synaptic connectivity data from the FAFB/FlyWire Drosophila brain connectome. Includes:

• Extracting and quantifying LC10a → AOTU019/AOTU025 → DNa02 synaptic pathways.
• Estimating visual receptive fields based on lobula input distributions.
• Mapping retinotopic organization by projecting medial-lateral lobula synapse positions to anterior-posterior retinal coordinates.
• Analyzing synapse counts, cable morphology, and connectivity strength across parallel visual-motor channels.

2. Data_Acquisition_Code/

Code used for behavioral and electrophysiological data collection. Includes:

• Integration with the G4 modular LED display system for open- and closed-loop visual stimulation.
• Real-time spherical treadmill tracking using FicTrac to measure forward, lateral, and rotational locomotion.
• Control of optogenetic arousal stimulation (P1 activation) and synchronization with behavioral output.
• Patch-clamp electrophysiology setup interfacing with visual stimuli and locomotion feedback.

3. Modeling_Code/

Mathematical models of pursuit behavior, including:

• A dynamical network model combining contralateral inhibition (AOTU019) and ipsilateral excitation (AOTU025).
• Simulation of pursuit behavior under varying error, object motion direction, arousal state, and forward velocity.
• Direction selectivity and gain scaling modules to test how circuit specializations improve pursuit performance.
• Variants that include random object motion, step perturbations, and genetic manipulations (e.g., AOTU019 silencing).
• Tools for quantifying model performance metrics such as settling time and midline fixation accuracy.

4. Data_Analysis_Code/

Data analysis pipelines for:

• Spike detection, membrane voltage filtering, and conversion of locomotor signals into kinematic variables.
• Binning and aligning data to visual stimuli for calculating average responses, tuning curves, and latencies.
• Computation of direction selectivity indices (DSIs), pursuit index metrics, and right-minus-left asymmetry signals.
• Classification of behavioral epochs (e.g., Schmitt triggers on forward velocity).
• Analysis of current injection experiments and cross-correlation between neural activity and rotational velocity.
• Mixed-effects statistical models for within-animal comparisons and between-genotype group analyses.

---

Requirements

• MATLAB 2023b+ for data acquisition and analysis
• R v4.4.2+ and RStudio v2024.12.0+ for connectome analyses
• Python 3.9+ for treadmill and display arena interfacing
• FicTrac for spherical treadmill tracking
• Display_Tools for modular LED display arena ("G4")

---

Citation

If you use this code or data, please cite:

Collie et al. (2025). Specialized parallel pathways for adaptive control of visual object pursuit. bioRxiv. https://doi.org/10.1101/2025.04.23.650240

---

Contact

For questions or data/code requests, please contact:
Rachel I. Wilson ([email protected])

---

© 2025 The Authors. Distributed under a CC BY-NC-ND 4.0 License.