An unsupervised learning approach for tracking mice in an enclosed area

Abstract

BackgroundIn neuroscience research, mouse models are valuable tools to understand the genetic mechanisms that advance evidence-based discovery. In this context, large-scale studies emphasize the need for automated high-throughput systems providing a reproducible behavioral assessment of mutant mice with only a minimum level of manual intervention. Basic element of such systems is a robust tracking algorithm. However, common tracking algorithms are either limited by too specific model assumptions or have to be trained in an elaborate preprocessing step, which drastically limits their applicability for behavioral analysis.ResultsWe present an unsupervised learning procedure that is basically built as a two-stage process to track mice in an enclosed area using shape matching and deformable segmentation models. The system is validated by comparing the tracking results with previously manually labeled landmarks in three setups with different environment, contrast and lighting conditions. Furthermore, we demonstrate that the system is able to automatically detect non-social and social behavior of interacting mice. The system demonstrates a high level of tracking accuracy and clearly outperforms the MiceProfiler, a recently proposed tracking software, which serves as benchmark for our experiments.ConclusionsThe proposed method shows promising potential to automate behavioral screening of mice and other animals. Therefore, it could substantially increase the experimental throughput in behavioral assessment automation.

Real-time analysis of the behaviour of groups of mice via a depth-sensing camera and machine learning

“…de Chaumont and colleagues (2012) published a semi-automatic software to describe precisely the types of interactions and the sequences of social events occurring between two visually indistinguishable mice, from a top-view video camera 36 , but this solution is limited as it need to be supervised and corrected by an expert. Unger and colleagues (2017) built on MiceProfiler and used another segmentation method to improve detection and avoid manual corrections 37 (manual interventions remained necessary at the beginning of the tracking). Hong and colleagues (2013) developed a system with two video cameras and one depth sensor to follow two mice 38 , but they needed to be of different coat colors to track the identities.…”

Mice Tracking Using The YOLO Algorithm

“…Commercial systems for video tracking of laboratory animals are often restricted to track the body or nose position, leading to inaccurate estimates of the head orientation or requiring manual artifices [Menezes et al 2018, Kretschmer et al 2012. Thus, large-scale studies emphasize the need for automated high-throughput systems providing a reproducible behavioral assessment of freely-moving mice with only a minimum level of manual intervention [Unger et al 2017].…”

“…Neuroscientists conduct a series of animal behavior experiments to validate their research and video tracking of animals is essential to achieve their goals [Menezes et al 2018, Unger et al 2017, Mathis et al 2018]. In the case of tracking video objects, animal tracking can be done in real-time, or even, in an offline analysis after the experiments; it will depend on the type of research being carried out.…”

Evaluation of caridean ecological distribution in the Ubatuba region, southeastern Brazilian coast using unsupervised machine learning technique

“…Unsupervised procedures identify the data structure entrance and, using a cluster approach, reveal the groups/categories of data (Hussain et al., 2017). Previous studies, in different areas, utilized such technique for the creation of ethograms of seabirds (Sakamoto et al., 2009), evaluation of locomotion of wild lions and dogs (Dewhirst et al., 2017), and social interactions of rats (Unger et al., 2017). Herein, we used this procedure for the first time to classify and grouping patterns of carideans' distribution.…”