Establishment and preliminary application of object recognition system based on DeepLabCut

This study aimed to develop a DeepLabCut (DLC)-based object recognition analysis system for assessing rodent cognitive function and validate its application in natural aging and elderly periodontitis mouse models. The system's hardware was constructed with a custom arena and high-definition industrial camera, and the DLC deep learning algorithm was trained to track five mouse body landmarks, enabling automatic quantification of 36 indicators across three categories: sniffing frequency, exploration duration, and novelty preference. The system subdivided exploratory behaviors by calibrating nose tip and body center, and set dynamic distance thresholds (1 cm, 1.5 cm, 2 cm) for the nose tip to capture fine-grained exploration. In the novel object recognition (NOR) and object location recognition (OLR) paradigms, traditional visual inspection failed to detect significant cognitive differences between young and aged mice, while the DLC system identified marked reductions in aged mice in the frequency and duration of body center and combined nose tip-body center exploration of the new object (2 cm away from the object), as well as corresponding novelty preference indices. In the elderly periodontitis models, traditional metrics showed increased nose tip exploration of the old object (2 cm away from the object) and reduced novelty preference in model mice; the DLC system further detected significantly elevated nose tip exploration frequency toward the old object (1.5 cm away from the object), accompanied by decreased frequency preference for exploration (1 cm away from the object). Collectively, this DLC-based system achieves sensitive, precise, and multidimensional quantification of mouse exploratory behavior, effectively distinguishing cognitive characteristics of aged and disease model mice. By overcoming the limitations of traditional methods, it captures subtle cognitive changes in aging and periodontitis models, screens key indicators for cognitive decline, and provides comprehensive behavioral evidence for elucidating the neural mechanisms underlying aging- and inflammation-associated cognitive impairment.