Abstract
Studying locomotor activity in animal models is crucial for understanding physiological, behavioral, and pathological processes. This study aimed to develop an artificial intelligence-based tracking system called Goblotrop, designed to localize rodents within their laboratory environment. The Goblotrop system uses two infrared cameras to record videos of rodents in their home cages. A neural network analyzes these videos to determine the rodent's position at each time point. By tracking changes in position over time, the system provides detailed insights into rodent behavior, including speed, mobility, and climbing activity. To evaluate the system's reliability, we utilized a starvation-induced hyperactivity model, employed as a female mouse model for anorexia nervosa. This model is characterized by pronounced hyperactivity, typically assessed using electronically monitored running wheels. Both the Goblotrop system and running wheel measurements demonstrated that starvation increases food-anticipatory activity (up to 4 h before food availability) while reducing nocturnal activity. The results from the Goblotrop system and running wheel measurements exhibited remarkable consistency. Thus, the Goblotrop system proves to be a valuable tool for studying locomotor activity and circadian rhythms in different cage areas in animal models. This tool provides potential for various scientific fields, including neuroscience, pharmacology, toxicology, and behavioral research.