Testing energy expenditures as biomarkers of aging using thermal imaging in a controlled environment

Project Abstract Rodent animal models are key to understanding the fundamental biological mechanism and to advancing clinically relevant testing, including aging and age-related diseases. Intimately linked to temperature, total energy expenditure (TEE) is an important concept to understand how the weight-reduced organismal state

improves metabolism or how physical activity affects energy expenditure. Intriguingly, an age-related increase in the energetic cost of physical activity in mice and age-related changes in physical activity energy expenditure (PAEE) in humans suggest the utility of these metrics as biomarkers of aging. We propose to

determine age-related changes in PAEE and/or other TEE constituents and their utility as physiological biomarkers of aging. Classical metabolic cages record only one mouse per cage, limiting the study of social behaviors, and commercial activity monitoring platforms do not offer continuous thermal imaging, are very

expensive, and require dedicated facilities and personnel. To address these shortcomings, we are developing a comprehensive platform for enduring thermal imaging in a controlled environment (ENTICE). Based on conventional home cages, ENTICE has thermal and far-red cameras, a controlled food dispenser, an in-cage

mouse weighing platform, and a volunteer running wheel with direct recording of energy generated by an animal. Thermal and far-red mages are analyzed on the fly, enabling real-time analysis of animal behaviors. Here, we propose to longitudinally record and analyze groups of young, middle-aged, and old mice to

determine age-related changes in PAEE and/or other constituent parts of TEE and to uncover common determinants and individual aspects (heterogeneity) of mouse aging. To enable the study of the social aspect of aging, we will develop a deep thermal identification (DeepT-ID) model that combines deep learning neural

networks with classical image texture features for rapid and robust thermal image analysis and mouse identification that enable distinguishing, simultaneous tracking, and analysis of several animals in the same cage. ENTICE represents a compact, affordable (under 10K USD) platform that could be installed and

operated in an individual laboratory. We provide a comprehensive assembly tutorial and a user-friendly Python-based software package to make behavioral studies accessible to all investigators working with mice. We believe that the development and validation of this novel thermal imaging-based scalable, cost-effective

platform will have a significant impact, enabling the testing of candidate drugs and interventions in various mouse models of aging and age-related diseases and capturing comprehensive behavioral and physiological phenotypes.