In preclinical research, the observation of non-human primate (NHP) behavior provides important data for neuroscience and drug development. Traditional behavior tracking methods require physical markers attached to animals. These markers can affect natural behavior, cause stress, and change movement patterns.
The AI-based NHP Behavior Analysis System (NBAS), also called BehaviorAtlas, was developed by Prisys Biotechnologies. The system uses markerless tracking, multi-view cameras, and deep learning algorithms to record and analyze primate behavior.
Markerless Technology for NHP Observation
NBAS uses markerless technology to capture 3D posture of non-human primates. The system does not require physical markers on the animals. This reduces discomfort and keeps the experimental process less invasive.
Researchers can observe natural movements without interference. This helps maintain normal behavior and improves the reliability of behavioral data. The approach also supports ethical standards in animal research.
Multi-View Synchronization for Movement Capture
NBAS records behavior with a multi-camera system. The system uses four or more high-speed cameras designed for 3D recording. The cameras are placed at different angles and record at the same time.
The system synchronizes the spatial coordinates and recording time of all cameras. This setup reduces visual occlusion and perspective error and allows accurate reconstruction of movement in three-dimensional space. The system can track motion with millimeter-level precision.
Deep Learning and 3D Skeleton Reconstruction
NBAS processes video data with deep learning algorithms. The system detects animal contours, body posture, and limb joints from the video recordings. The algorithms work under different lighting conditions and maintain stable detection performance.
The software reconstructs a 3D skeleton model of the animal. It tracks more than 21 key body points and records their position during movement.
The system then analyzes behavior sequences automatically. It divides continuous motion into behavior segments and estimates behavioral states with second-level temporal resolution. This automated analysis reduces manual work and lowers the risk of human error in data extraction.
Applications in Neuroscience and Drug Development
NBAS provides objective behavior measurements for neuroscience research. The system is used to analyze behavioral patterns in several neurological disease models, including:
- Parkinson's disease models
- Pain models
- Middle cerebral artery occlusion (MCAO) stroke models
- Epilepsy models
The system also supports drug development studies. Researchers can measure joint movement and body motion patterns during treatment studies. These measurements help evaluate the behavioral effects of pharmacological interventions.
The collected data provides quantitative evidence that supports preclinical evaluation and later clinical study design. Through this system, Prisys Biotechnologies provides automated tools for animal behavior analysis in translational research.
