Automatic Detection of Stress Indicators in Laboratory Animals Based on Video Recordings

 

Motivation

Many types of animal experiments need to be assessed in terms of severity, where the burden of the experiment on the animals is classified into different levels. This determination of the burden placed on animals by experiments (severity assessment) is an essential aspect of animal experimentation. In order to develop, validate and implement new, evidence-based methods for this purpose, the DFG research group FOR 2591 has been founded. Within this group, the LfB performs tasks in image and video based analysis of data.

Scientific Questions

The aim of the project is to develop and validate markers for animal stress. In addition to a variety of clinical markers (laboratory values, vital signs), visual assessment is playing an increasingly important role. One example is the Mouse Grimace Scale (MGS) method, in which changes in the facial features of mice allow conclusions to be drawn about their stress. Visual assessment has significant advantages, including that it can be performed with simple training in everyday life, does not require special measuring equipment, and is noninvasive. These advantages are offset by the high time required for collection and reporting of findings. In addition, visual findings are partially dependent on the person making the assessment and are therefore subjective. In order to eliminate these problems, the LfB is developing an automatic method for determining the MGS as part of the project. Neural networks for animal detection, behavior analysis and MGS determination will be developed and trained with images judged by experts. The goal is fully automated MGS classification without significant human assistance.

Video-based behavioral analysis in the home cage is also being explored as a major focus. The home cage is the cage that the animals inhabit during the experiments. Here, behavior can be observed without interference from humans or experiments. For this purpose, a special cage attachment is being developed within the research group, which is equipped with a variety of cameras in the visible, near-infrared and thermal spectral range. This system will be used to detect and track the animals, to study their social behavior and to examine various vital parameters such as heart and respiration rate. The LfB is responsible for detection, identification, tracking, social analysis and MGS scoring in the cages. Here, too, current developments in machine learning are used in the methodological development in order to be able to analyze the complex data reliably.

Partners

External Funding

  • DFG Research Unit FOR 2591, “Severity assessment in animal-based research”, Project nr. 321137804
  • DFG Research Grant, “Automated measurement of stress scores in video recordings of laboratory mice”; Project nr. 408132301

Kontakt

Publications

2022

Marcin Kopaczka, Lisa Ernst, Mareike Schulz, Rene Tolba and Dorit Merhof
Computation of Traveled Distance of Pigs in an Open Field With Fully Convolutional Neural Networks
In: Bildverarbeitung fuer die Medizin (BVM)

2022

Eva Zentrich, Laura Wassermann, Birgitta Struve, Kristin Selke, Manuela Buettner, Lydia Maria Keubler, Janin Reifenrath, Nina Angrisani, Merle Kempfert, Annika Krause, Olaf Bellmann, Marcin Kopaczka, Dorit Merhof, Marion Bankstahl, André Bleich and Christine Häger
Post-operative severity assessment in sheep

2022

Lisa Ernst, Stefan Bruch, Marcin Kopaczka, Dorit Merhof, Andre Bleich, Rene H. Tolba, Steven R. Talbot
A Model-Specific Simplification of the Mouse Grimace Scale Based on the Pain Response of Intraperitoneal CCl4 Injections
In: Scientific Reports

2020

Marcin Kopaczka, Tobias Jacob, Lisa Ernst, Mareike Schulz, Rene Tolba and Dorit Merhof
Robust Open Field Rodent Tracking using a Fully Convolutional Network and a Softargmax Distance Loss
In: Bildverarbeitung fuer die Medizin (BVM)

2019

Lisa Ernst, Marcin Kopaczka, Mareike Schulz, Steven R. Talbot, Leonie Zieglowski, Marco Meyer, Stefan Bruch, Dorit Merhof and René H. Tolba
Improvement of the Mouse Grimace Scale set-up for Implementing of a Semi-automated Mouse Grimace Scale MGS Scoring (Part 1)
In: Laboratory animals

2019

Lisa Ernst, Marcin Kopaczka, Mareike Schulz, Steven R. Talbot, Birgitta Struve, Christine Häger, André Bleich, Mattea Durst, Paulin Jirkof, Margarete Arras, Roelof Maarten van Dijk, Nina Miljanovic, Heidrun Potschka, Dorit Merhof and René H. Tolba
Semi-automated generation of pictures for the Mouse Grimace Scale: A multi-laboratory analysis (Part 2)
In: Laboratory animals

2019

Marcin Kopaczka, Daniel Tillmann, Lisa Ernst, Justus Schock, Ren'e H. Tolba and Dorit Merhof
Assessment of Laboratory Mouse Activity in Video Recordings Using Deep Learning Methods
In: International Engineering in Medicine and Biology Conference (EMBC)

2018

Marcin Kopaczka, Lisa Ernst, Jakob Heckelmann, Christoph Schorn, René Tolba and Dorit Merhof
Automatic Key Frame Extraction From Videos For Efficient Mouse Pain Scoring
In: 5th International Conference on Signal Processing and Integrated Networks (SPIN)