The iMouse System – A Visual Method for Standardized Digital Data Acquisition Reduces Severity Levels in Animal-Based Studies

Author(s): Maciej Laz, Mirko Lampe, Isaac Connor, Dmytro Shestachuk, Marcel Ludwig, Ursula Müller, Oliver F. Strauch, Nadine Suendermann, Stefan Lüth, Janine Kah

In translational research, using experimental animals remains the preferred standard for assessing the effectiveness of potential therapeutic interventions, particularly when investigating physiological interactions and relationships. However, the execution of these investigations is contingent upon minimizing the impact on the well-being of the experimental animals. To evaluate the severity level of the animals, inspections were conducted routine observations, multiple times each day, visually. It is noted that these visual assessments disrupt the animals during their periods of rest, resulting in elevated stress levels, which, in turn, exacerbate the animals' burden and may consequently exert an influence on the outcomes of scientific studies. Our study examined the feasibility of implementing a digital monitoring system in a translational study conducted within IVC cages. Our objective was to determine whether a camera-based observation system could reduce manual visual inspections and whether digitally available data from this study could be utilized to train an algorithm capable of distinguishing between activities like drinking. Furthermore, we aimed to ascertain whether the system could monitor the recovery phase following experimentally induced high stress, potentially as a substitute for frequent visual inspections. Within the scope of our study, we successfully demonstrated the feasibility of integrating iMouse system hardware components into the existing IVC (Individually Ventilated Cage) racks. Importantly, we established that this system can be accessed remotely from outside the animal facility, thus facilitating comprehensive digital surveillance of the experimental subjects. Furthermore, the digital biomarkers (digitally acquired data out of the home cage) proved instrumental in training algorithms capable of analyzing the long-term drinking behavior of the animals. In summary, our work has yielded an integrated, retrofittable, and modular system that serves two critical criteria. Firstly, it enables the execution of visual inspections without disturbing the animals. Secondly, it enhances the traceability and transparency of research involving animal subjects employing digital data capture by generating digital biomarkers.