Title
Teaching AI to Read Mouse Behavior
Leaders
Amy Conner (@amynconner19)
Megan Guidry (@meganmguidry)
Megan Beers Wood
Project description
The Behavioral Inventory of Mouse Affective Pain (BIOMAP) is a machine learning–based behavioral assay developed in the HARMONIC Laboratory at Vanderbilt University Medical Center. BIOMAP combines DeepLabCut, SimBA, and custom Python analyses to quantify pain-related behavior from mouse videos.
Our goal is to transform the current multi-step workflow into an open-source, automated pipeline that can process batches of behavioral videos with minimal user intervention. The pipeline will automatically run trained DeepLabCut models, execute the existing BIOMAP and SimBA analyses, calculate baseline-normalized behavioral metrics and composite pain scores, and generate publication-ready figures and analysis-ready datasets.
As a secondary objective, we are interested in improving behavioral-state classification by distinguishing pain-related immobility from sleep, rest, and quiet wakefulness using pose-estimation data and machine-learning approaches.
This project is exciting because it combines neuroscience, computer vision, machine learning, and scientific software engineering to solve a real research problem. Contributors can choose from a wide range of tasks, from pipeline automation and visualization to behavioral classification and documentation, and their work has the potential to become part of an open-source tool that will continue to be developed and used by researchers beyond BrainHack.
Link to project repository/sources
Project Github Link
Concerete goals with specific tasks for Brainhack Vanderbilt 2026
1. Automated BIOMAP Pipeline
Develop an end-to-end, open-source pipeline that automatically processes batches of behavioral videos from raw video to publication-ready results.
2. Behavioral-State Classification
Develop methods to distinguish pain-related immobility from sleep, rest, and other low-movement behavioral states.
🟢 Good first issues
- Design a standardized metadata schema for BIOMAP behavioral videos (Issue #1)
- Automate baseline normalization for each sound-level condition (Issue #4)
- Automate Facial Grimace composite score calculation (Issue #6)
- Automate Body Position composite score calculation (Issue #7)
- Export standardized, analysis-ready datasets (Issue #10)
Skills
We welcome contributors with a variety of backgrounds, including:
- Python: scripting, automation, data processing
- Machine learning & computer vision: DeepLabCut, SimBA (experience helpful, not required)
- Data visualization: matplotlib, Plotly, figure generation
- Neuroscience: behavioral analysis, pain research, experimental design
- Git & GitHub: version control, documentation, open-source collaboration
Tasks are available for beginner, intermediate, and advanced contributors.
Onboarding documentation
📖 README
What will participants learn?
Participants will gain experience in:
- Scientific software development in Python
- Computer vision and machine learning (DeepLabCut, SimBA)
- Behavioral data analysis and visualization
- Building automated, reproducible analysis pipelines
- Quantitative behavioral neuroscience
- Open-source collaboration with Git and GitHub
Public data to use
We will provide example data containing representative behavioral videos, DeepLabCut tracking outputs (CSV files), and example analysis outputs for pipeline development and testing.
Additional background and methodology are available in the BIOMAP preprint
Number of collaborators
4+
Credit to collaborators
Meaningful scientific and software contributions will be recognized in future software releases, presentations, preprints, and publications, as appropriate and in accordance with standard scientific authorship and contribution practices.
Image
Project Summary
Build an open-source pipeline that uses AI to automatically analyze mouse behavior and distinguish pain-related behaviors from other behavioral states.
Type
pipeline_development
Development status
1_basic structure
Topic
machine_learning
Tools
other
Programming language
Python
Modalities
behavioral
Git skills
1_commit_push
Anything else?
No response
Things to do after the project is submitted and ready to review.
Title
Teaching AI to Read Mouse Behavior
Leaders
Amy Conner (@amynconner19)
Megan Guidry (@meganmguidry)
Megan Beers Wood
Project description
The Behavioral Inventory of Mouse Affective Pain (BIOMAP) is a machine learning–based behavioral assay developed in the HARMONIC Laboratory at Vanderbilt University Medical Center. BIOMAP combines DeepLabCut, SimBA, and custom Python analyses to quantify pain-related behavior from mouse videos.
Our goal is to transform the current multi-step workflow into an open-source, automated pipeline that can process batches of behavioral videos with minimal user intervention. The pipeline will automatically run trained DeepLabCut models, execute the existing BIOMAP and SimBA analyses, calculate baseline-normalized behavioral metrics and composite pain scores, and generate publication-ready figures and analysis-ready datasets.
As a secondary objective, we are interested in improving behavioral-state classification by distinguishing pain-related immobility from sleep, rest, and quiet wakefulness using pose-estimation data and machine-learning approaches.
This project is exciting because it combines neuroscience, computer vision, machine learning, and scientific software engineering to solve a real research problem. Contributors can choose from a wide range of tasks, from pipeline automation and visualization to behavioral classification and documentation, and their work has the potential to become part of an open-source tool that will continue to be developed and used by researchers beyond BrainHack.
Link to project repository/sources
Project Github Link
Concerete goals with specific tasks for Brainhack Vanderbilt 2026
1. Automated BIOMAP Pipeline
Develop an end-to-end, open-source pipeline that automatically processes batches of behavioral videos from raw video to publication-ready results.
2. Behavioral-State Classification
Develop methods to distinguish pain-related immobility from sleep, rest, and other low-movement behavioral states.
🟢 Good first issues
Skills
We welcome contributors with a variety of backgrounds, including:
Tasks are available for beginner, intermediate, and advanced contributors.
Onboarding documentation
📖 README
What will participants learn?
Participants will gain experience in:
Public data to use
We will provide example data containing representative behavioral videos, DeepLabCut tracking outputs (CSV files), and example analysis outputs for pipeline development and testing.
Additional background and methodology are available in the BIOMAP preprint
Number of collaborators
4+
Credit to collaborators
Meaningful scientific and software contributions will be recognized in future software releases, presentations, preprints, and publications, as appropriate and in accordance with standard scientific authorship and contribution practices.
Image
Project Summary
Build an open-source pipeline that uses AI to automatically analyze mouse behavior and distinguish pain-related behaviors from other behavioral states.
Type
pipeline_development
Development status
1_basic structure
Topic
machine_learning
Tools
other
Programming language
Python
Modalities
behavioral
Git skills
1_commit_push
Anything else?
No response
Things to do after the project is submitted and ready to review.
Hi @brainhack-vandy/project-monitors my project is ready!