sociSleep

Identity-preserving tracking of individual Drosophila behavior in social environments.

sociSleep is an identity-preserving behavioral tracking system for quantifying sleep, locomotion, and social dynamics of individual Drosophila housed in social environments.

Unlike conventional Drosophila sleep systems that require physical isolation of animals, sociSleep continuously tracks individuals within shared arenas, allowing long-term measurement of sleep while preserving social interactions.

---

Features

• Identity-preserving tracking for individual flies
• Quantification of sleep and locomotor activity
• Simultaneous tracking of social groups
• Automatic detection of merged flies
• Support for multiple experimental plate layouts
• Multi-camera acquisition (1–2 USB cameras)
• Independent output for each camera
• Interactive OpenCV graphical interface
• Automatic CSV data logging

---

Experimental Designs

Currently supported plate configurations:

Design	Left Arena	Right Arena
Solo vs Solo	1 fly	1 fly
Solo vs Group	1 fly	2 flies
Group vs Group	2 flies	2 flies

Additional designs can be added by modifying plate_design.py.

---

System Architecture

USB Camera(s)
        │
        ▼
 FlyTracker Thread(s)
        │
        ▼
 Identity Tracking
        │
        ▼
 Behavior Detection
        │
        ▼
 CSV Logger

Each USB camera is handled by an independent tracking thread.

The dashboard combines all active cameras into a single live display.

Example:

+----------------------+----------------------+
|     Camera 1         |      Camera 2        |
|                      |                      |
|   Arena A  Arena B   |   Arena A  Arena B   |
|                      |                      |
+----------------------+----------------------+

---

Directory Structure

sociSleep/

├── main.py
├── fly_tracker.py
├── dashboard.py
├── camera_detector.py
├── arena_config.py
├── design_selector.py
├── csv_logger.py
├── plate_design.py
├── config.py
└── ...

---

Requirements

• Python 3.11+
• OpenCV-python=4.5.0
• NumPy
• pygrabber
• SciPy
• Pandas

Install dependencies:

pip install -r requirements.txt

---

Running sociSleep

camera_names.py

The program automatically detects connected USB cameras, and output camera names.

camera_detector.py

Modify camera names (just keyword) at Line 7.

run_sociSleep.py

Launch sociSleep tracker.

---

Multi-camera Support

The current version supports one or two USB cameras simultaneously.

Each camera operates independently while sharing the same experimental plate design.

Output files are generated separately:

results_camera1/
results_camera2/

Each directory contains:

• Tracking CSV
• Experiment log
• Optional snapshots

---

Plate Design Selection

Before tracking begins, select one of the available plate layouts:

• Solo vs Solo
• Solo vs Group
• Group vs Group

The graphical control panel provides:

• Large clickable buttons
• Current design highlight
• Arena summary
• READY / LOCKED status

Once tracking starts, the plate design becomes locked to prevent accidental changes.

---

Output

Each experiment generates CSV files containing:

• Timestamp
• Fly coordinates
• Fly movement (0:immobile; 1: moving)
• Arena assignment

Example:

Time	A_x	A_y	A_movement	B_x	B_y	B_movement

---

Analysis

Integrated sociSleep_analyzer: version 4.3

---

Identity Tracking

sociSleep preserves fly identity during long-term recordings using:

• Kalman filter + nearest assignment
• Adaptative weight of size-based identity assignment
• Merge detection
• Arena-specific constraints

Identity switches are minimized during brief social interactions.

---

GUI

The OpenCV interface consists of:

Dashboard

• Live camera feeds
• Arena overlays
• Fly identities
• Tracking status

Plate Design

• Clickable plate layout selection
• Arena information
• Tracking status

Arena Calibration

Each camera has an independent calibration window for selecting arena positions before tracking.

---

Typical Workflow

1. Connect USB camera(s).
2. Launch run_sociSleep.py.
3. Select the experimental plate design.
4. Calibrate arena locations and camera parameters.
5. Press START.
6. Monitor tracking in the live dashboard.
7. CSV files are automatically saved during acquisition.

---

Citation

If you use sociSleep in your research, please cite:

Binbin Wu, William Ja. Rencent social loss, not chronic isolation, reshapes sleep in Drosophila. Doi: https://doi.org/10.64898/2026.07.12.738068

---

License

GPL-3.0 open-source License

---

Contact

Binbin Wu

The Wertheim UF Scripps Institute

For questions, bug reports, or feature requests, please open a GitHub Issue.