Biology subjects

Hellwich, O.

Publications and source records attributed to Hellwich, O..

6 recordsLinked to original sources

bioRxiv2026

Different Paradigms from Computer Vision Align with Human Assessment of the Mouse Grimace Scale

Animal welfare is a central aspect in animal-based research where mice are most commonly used. Their facial expression can be analyzed to assess their well-being status using the Mouse Grimace Scale. However, its manual application becomes increasingly impractical when used on a large number of animals. This lead to the ongoing integration of computer vision methods to automate the analysis. While such methods have proven effective qualitatively, a systematic assessment to verify their reliability largely remains an open research gap. In this work, we attempted to close this gap as we evaluated three dominant paradigms (i.e., classification from supervised learning features, self-supervised learning features, or landmark locations) for the binary (i.e., well-being un-/impaired) classification of facial mouse images. Our quantitative results showed that such methods can be employed successfully with as low as 16% type II error rates. For qualitative assessment, we visualized the decision-making process and demonstrated that mainly pixels associated with the mouse rather than its environment are used. We further discovered that visual characteristics of the mice beyond those described by the Mouse Grimace Scale contributed to the classification. Our work showed that the automated well-being status assessment in mice is trustworthy and urges towards widespread adoption.

animal behavior and cognition↗

bioRxiv2026

Automatic pain face analysis in mice: Applied to a varied dataset with non-standardized conditions

Biomedical research relies on scientifically validated tools to assess pain, suffering, and distress in laboratory animals to ensure their well-being. In mice, the most frequently used laboratory animals, the Mouse Grimace Scale (MGS) provides a reliable tool for the assessment of facial expression changes caused by impaired well-being. However, no automated tool can yet reliably assess all features of the MGS across different mouse strains under varying experimental or housing conditions in real-time, as the variability present in recorded image datasets poses substantial challenges for computer vision models. Despite this technical difficulty, variability across subsets in terms of mouse strain, treatments, laboratory, and image acquisition setup is essential for paving the way toward MGS assessment under non-standardized conditions in the home cage rather than standardized cage-side recording setups. Against this background, a large and diverse dataset containing five subsets is introduced and a deep learning model was trained to predict the average MGS scores ranging between 0 and 2. It achieved a root mean squared error (RMSE) of 0.26 when trained on all subsets of the dataset, outperforming the average human rater in terms of error magnitude. The correlation between human raters and automated MGS scores was very high (Pearsons r=0.85). In the cross-dataset evaluation, one subset was excluded from training and used for testing the model. This approach yielded higher errors compared to models trained and tested on the same subsets. A model restricted to the feature of orbital tightening showed lower performance than one trained on all facial features of the MGS. Overall, the most reliable model for predicting average MGS scores for a novel dataset is the one trained on the combined subsets. Performance may be further enhanced by fine-tuning the model using human-generated MGS scores for a portion of the novel subset.

animal behavior and cognition↗

bioRxiv2025

Towards Fully Automated Investigation of Social Learning in Mice

Mice have been demonstrated to learn from each other in social interactions, the extent to which this takes place and the strategies involved, however, largely remain to be elucidated beyond spatially and temporally confined tests of social memory retention. Here, we present a method which utilizes and modifies 1) a commercially available tool for automated behavioral testing, the IntelliCage and 2) an open-source solution for 24/7 live animal tracking, the Live Mouse Tracker, to create a powerful method for the investigation of learning behavior in semi-naturalistic group settings. We see a wide range of possible applications, such as for instance the investigation of learning in social interactions, which we present here as an example. In the present study, co-learning did not facilitate place learning over individual learning. While automated annotation of behaviors was effective, markerless animal identification proved unreliable in a highly enriched environment with manifold opportunity of occlusion from video tracking. In response, we here present a rationale for identifying the reliable portion of tracking data, to which we confine the behavioral analysis. Co-learning animals engaged more often in some prosocial interactions with their teammates than with other animals, but they did overall not interact with each other more frequently than with individual learners. Correlative analysis of learning behavior and social interactions did not reveal any particular association between behavior and learning success. While the mechanisms of social learning in mice could not be conclusively elucidated within the scope of this study, we report on the development of a promising tool for presenting manifold learning tasks to mice while tracking their individual and social behaviors in a fully automated manner. Further, we discuss limitations of the current configuration and present an outlook on further improving the method.

animal behavior and cognition↗

bioRxiv2024

Mechanical problem solving in mice

Behavioral learning is a complex phenomenon that involves cognitive, perceptual, motor and emotional contributions. Yet, most animal studies focus on reductionist tasks that aim to isolate one of these aspects [1, 2, 3, 4]. Here, we use a lockbox -- a complex, multi-step mechanical puzzle -- to study learning dynamics in freely behaving mice. The mice engaged spontaneously with the task and learned to solve it within just a few trials. To dissect different contributions to this rapid form of learning, we combined deep learning-based behavioral tracking in a multi-camera setup with probabilistic inference and computational modeling. We find that the learning progress of the mice was initially dominated by the acquisition of motor skills, i.e., the increased ability to manipulate the individual mechanisms, while a cognitive strategy for the task sequence emerged only later. The lockbox paradigm may hence offer a promising framework for studying the interaction between low-level motor learning and high-level decision-making strategies in a single, ethologically relevant task.

animal behavior and cognition↗

bioRxiv2023

Objects Guide Human Gaze Behavior in Dynamic Real-World Scenes

The complexity of natural scenes makes it challenging to experimentally study the mechanisms behind human gaze behavior when viewing dynamic environments. Historically, eye movements were believed to be driven primarily by space-based attention towards locations with salient features. Increasing evidence suggests, however, that visual attention does not select locations with high saliency but operates on attentional units given by the objects in the scene. We present a new computational framework to investigate the importance of objects for attentional guidance. This framework is designed to simulate realistic scanpaths for dynamic real-world scenes, including saccade timing and smooth pursuit behavior. Individual model components are based on psychophysically uncovered mechanisms of visual attention and saccadic decision-making. All mechanisms are implemented in a modular fashion with a small number of well-interpretable parameters. To systematically analyze the importance of objects in guiding gaze behavior, we implemented five different models within this framework: two purely spatial models, where one is based on low-level saliency and one on high-level saliency, two object-based models, with one incorporating low-level saliency for each object and the other one not using any saliency information, and a mixed model with object-based attention and selection but space-based inhibition of return. We optimized each models parameters to reproduce the saccade amplitude and fixation duration distributions of human scanpaths using evolutionary algorithms. We compared model performance with respect to spatial and temporal fixation behavior, including the proportion of fixations exploring the background, as well as detecting, inspecting, and returning to objects. A model with object-based attention and inhibition, which uses saliency information to prioritize between objects for saccadic selection, leads to scanpath statistics with the highest similarity to the human data. This demonstrates that scanpath models benefit from object-based attention and selection, suggesting that object-level attentional units play an important role in guiding attentional processing. Author summaryThere has long been an interest in understanding how we decide when and where to move our eyes, and psychophysical experiments have uncovered many underlying mechanisms. Under controlled laboratory conditions, objects in the scene play an important role in guiding our attention. Due to the visual complexity of the world around us, however, it is hard to assess experimentally how objects influence eye movements when observing dynamic real-world scenes. Computational models have proved to be a powerful tool for investigating visual attention, but existing models are either only applicable to images or restricted to predicting where humans look on average. Here, we present a computational framework for simulating where and when humans decide to move their eyes when observing dynamic real-world scenes. Using our framework, we can assess the influence of objects on the model predictions. We find that including object-based attention in the modeling increases the resemblance of simulated eye movements to human gaze behavior, showing that objects play indeed an important role in guiding our gaze when exploring the world around us. We hope that the availability of this framework encourages more research on attention in dynamic real-world scenes.

neuroscience↗

bioRxiv2019

Towards a fully automated surveillance of well-being status in laboratory mice using deep learning

Assessing the well-being of an animal is hindered by the limitations of efficient communication between humans and animals. Instead of direct communication, a variety of behavioral, biochemical, physiological, and physical parameters are employed to evaluate the well-being of an animal. Especially in the field of biomedical research, scientifically sound tools to assess pain, suffering, and distress for experimental animals are highly demanded due to ethical and legal reasons. For mice, the most commonly used laboratory animals, a valuable tool is the Mouse Grimace Scale (MGS), a coding system for facial expressions of pain in mice which has been shown to be accurate and reliable. Currently, MGS scoring is very time and effort consuming as it is manually performed by humans being thoroughly trained in using this method. Therefore, we aim to develop a fully automated system for the surveillance of well-being in mice. Our work introduces a semi-automated pipeline as a first step towards this goal. We use and provide a new data set of images of black-furred laboratory mice that were moving freely, thus the images contain natural variation with regard to perspective and background. The analysis of this data set is therefore more challenging but reflects realistic conditions as it would be obtainable without human intervention. Images were obtained after anesthesia (with isoflurane or ketamine/xylazine combination) and surgery (castration). We deploy two pre-trained state of the art deep convolutional neural network (CNN) architectures (ResNet50 and InceptionV3) and compare to a third CNN architecture without pre-training. Depending on the particular treatment, we achieve an accuracy of up to 99% for binary \"pain\"/\"no-pain\" classification.\n\nAuthor summaryIn the field of animal research, it is crucial to assess the well-being of an animal. For mice, the most commonly used laboratory animals, there is a variety of indicators for well-being. Especially the facial expression of a mouse can give us important information on its well-being state. However, currently the surveillance of well-being can only be ensured if a human is present. Therefore, we developed a first approach towards a fully automated surveillance of the well-being status of a mouse. We trained neural networks on face images of black-furred mice, which were either untreated or underwent anesthesia or surgery, to distinguish between an impaired and unimpaired well-being state. Our systems successfully learnt to assess whether the well-being of a mouse was impaired and, depending on the particular treatment, its decision was correct in up to 99%. A tool that visualizes the features used for the decision making process indicated that the decision was mainly based on the facial expressions of a mouse.

animal behavior and cognition↗