BehaviorDEPOT is a simple, flexible tool for automated behavioral detection based on markerless pose tracking.

Area of Science:

• Neuroscience
• Ethology
• Computational Biology

Background:

• Quantitative analysis of animal behavior is crucial for understanding neural mechanisms of cognition and emotion.
• Manual or commercial software-based behavioral analysis is often labor-intensive, costly, and inflexible.
• Markerless pose estimation facilitates automated animal tracking but behavior classification remains challenging.

Purpose of the Study:

• To develop a user-friendly and flexible software tool, BehaviorDEPOT, for automated behavioral analysis from video data.
• To enable researchers, even those without programming skills, to detect and classify specific animal behaviors.
• To facilitate the integration of behavioral data with neural recordings.

Main Methods:

• BehaviorDEPOT utilizes keypoint tracking data to calculate kinematic and postural statistics.
• The software employs user-defined or pre-programmed heuristics to reliably detect specific behaviors.
• It offers a graphical interface for heuristic development and incorporates framewise data storage for neural data alignment.

Main Results:

• BehaviorDEPOT accurately detects behaviors from video timeseries, applicable to diverse experimental designs.
• A hard-coded heuristic for freezing detection achieved over 90% accuracy in mice and rats.
• The software demonstrated utility in various assays, including fear conditioning, T-maze decision-making, and open field tests.

Conclusions:

• BehaviorDEPOT provides a simple, flexible, and accessible solution for automated animal behavior analysis.
• The software empowers researchers to analyze complex behaviors and integrate with neural data, advancing cognitive and emotional process studies.
• Its no-code interface and broad applicability make it a valuable tool for neuroscience laboratories.