Does advancement in marker-less pose-estimation mean more quality research? A systematic review

Area of Science:

• Computer Vision
• Neuroscience
• Behavioral Science

Background:

• Marker-less pose estimation has revolutionized computer vision.
• Its application in scientific research, particularly rodent models, needs systematic documentation.
• Despite advanced algorithms, current usage and integration in studies remain unclear.

Purpose of the Study:

• To systematically review the application of pose-estimation techniques in rodent models.
• To identify trends, gaps, and patterns in the use of these technologies.
• To assess the adoption rate and integration of marker-less pose estimation in scientific research.

Main Methods:

• Comprehensive literature search across Crossref, OpenAlex, PubMed, and Scopus (2016-2025).
• AI-assisted screening of ~1,000 titles and abstracts from 16,412 retrieved entries.
• Categorization of 67 selected papers into tool-focused, method-focused, and study-focused studies.

Main Results:

• Publication frequency of rodent pose-estimation studies has accelerated, with over half published post-2021.
• 30 tool-focused, 28 method-focused, and 9 study-focused papers were identified.
• Emerging trends include new keypoint-detection methods and integration into behavioral experiments across disease models.

Conclusions:

• Marker-less pose-estimation technologies show significant progress but face limited widespread application.
• Traditional behavioral assays are still prevalent, indicating under-utilization of advanced tools.
• Standardized protocols are essential to bridge the gap and unlock the full potential of these technologies for preclinical behavioral science.