Most downloaded biology preprints, all time
in category animal behavior and cognition
1,933 results found. For more information, click each entry to expand.
35,958 downloads bioRxiv animal behavior and cognition
This tutorial introduces the reader to Gaussian process regression as an expressive tool to model, actively explore and exploit unknown functions. Gaussian process regression is a powerful, non-parametric Bayesian approach towards regression problems that can be utilized in exploration and exploitation scenarios. This tutorial aims to provide an accessible introduction to these techniques. We will introduce Gaussian processes which generate distributions over functions used for Bayesian non-parametric regression, and demonstrate their use in applications and didactic examples including simple regression problems, a demonstration of kernel-encoded prior assumptions and compositions, a pure exploration scenario within an optimal design framework, and a bandit-like exploration-exploitation scenario where the goal is to recommend movies. Beyond that, we describe a situation modelling risk-averse exploration in which an additional constraint (not to sample below a certain threshold) needs to be accounted for. Lastly, we summarize recent psychological experiments utilizing Gaussian processes. Software and literature pointers are also provided.
18,834 downloads bioRxiv animal behavior and cognition
Dominique Grandjean, Riad Sarkis, Jean-Pierre Tourtier, Clothilde Julien-Lecocq, Aymeric Benard, Vinciane Roger, Eric Levesque, Eric Bernes-Luciani, Bruno Maestracci, Pascal Morvan, Eric Gully, David Berceau-Falancourt, Jean-Luc Pesce, Bernard Lecomte, Pierre Haufstater, Gregory Herin, Joaquin Cabrera, Quentin Muzzin, Capucine Gallet, Hélène Bacqué, Jean-Marie Broc, Leo Thomas, Anthony Lichaa, Georges Moujaes, Michele Saliba, Aurore Kuhn, Mathilde Galey, Benoit Berthail, Lucien Lapeyre, Olivier Méreau, Marie-Nicolas Matteï, Audrey Foata, Louisa Bey, Anne-Sophie Philippe, Paul Abassi, Ferri Pisani, Marlène Delarbre, Jean-Marc Orsini, Anthoni Capelli, Steevens Renault, Karim Bachir, Anthony Kovinger, Eric Comas, Aymeric Stainmesse, Erwan Etienne, Sébastien Voeltzel, Sofiane Mansouri, Marlène Berceau-Falancourt, Brice Leva, Frederic Faure, Aimé Dami, Marc Antoine Costa, Jean-Jacques Tafanelli, Jean-Benoit Luciani, Jean-Jacques Casalot, Lary Charlet, Eric Ruau, Mario Issa, Carine Grenet, Christophe Billy, Loic Desquilbet
The aim of this study is to evaluate if the sweat produced by COVID-19 persons (SARS-CoV-2 PCR positive) has a different odour for trained detection dogs than the sweat produced by non COVID-19 persons. The study was conducted on 3 sites, following the same protocol procedures, and involved a total of 18 dogs. A total of 198 armpits sweat samples were obtained from different hospitals. For each involved dog, the acquisition of the specific odour of COVID-19 sweat samples required from one to four hours, with an amount of positive samples sniffing ranging from four to ten. For this proof of concept, we kept 8 dogs of the initial group (explosive detection dogs and colon cancer detection dogs), who performed a total of 368 trials, and will include the other dogs in our future studies as their adaptation to samples scenting takes more time. The percentages of success of the dogs to find the positive sample in a line containing several other negative samples or mocks (2 to 6) were 100p100 for 4 dogs, and respectively 83p100, 84p100, 90p100 and 94p100 for the others, all significantly different from the percentage of success that would be obtained by chance alone. We conclude that there is a very high evidence that the armpits sweat odour of COVID-19+ persons is different, and that dogs can detect a person infected by the SARS-CoV-2 virus. ### Competing Interest Statement The authors have declared no competing interest.
10,909 downloads bioRxiv animal behavior and cognition
Variation across dog breeds presents a unique opportunity for investigating the evolution and biological basis of complex behavioral traits. We integrated behavioral data from more than 17,000 dogs from 101 breeds with breed-averaged genotypic data (N = 5,697 dogs) from over 100,000 loci in the dog genome. Across 14 traits, we found that breed differences in behavior are highly heritable, and that clustering of breeds based on behavior accurately recapitulates genetic relationships. We identify 131 single nucleotide polymorphisms associated with breed differences in behavior, which are found in genes that are highly expressed in the brain and enriched for neurobiological functions and developmental processes. Our results provide insight into the heritability and genetic architecture of complex behavioral traits, and suggest that dogs provide a powerful model for these questions.
8,777 downloads bioRxiv animal behavior and cognition
Recent work quantifying postural dynamics has attempted to define the repertoire of behaviors performed by an animal. However, a major drawback to these techniques has been their reliance on dimensionality reduction of images which destroys information about which parts of the body are used in each behavior. To address this issue, we introduce a deep learning-based method for pose estimation, LEAP (LEAP Estimates Animal Pose). LEAP automatically predicts the positions of animal body parts using a deep convolutional neural network with as little as 10 frames of labeled data for training. This framework consists of a graphical interface for interactive labeling of body parts and software for training the network and fast prediction on new data (1 hr to train, 185 Hz predictions). We validate LEAP using videos of freely behaving fruit flies (Drosophila melanogaster) and track 32 distinct points on the body to fully describe the pose of the head, body, wings, and legs with an error rate of <3% of the animal's body length. We recapitulate a number of reported findings on insect gait dynamics and show LEAP's applicability as the first step in unsupervised behavioral classification. Finally, we extend the method to more challenging imaging situations (pairs of flies moving on a mesh-like background) and movies from freely moving mice (Mus musculus) where we track the full conformation of the head, body, and limbs.
7,630 downloads bioRxiv animal behavior and cognition
Fabrice de Chaumont, Elodie Ey, Nicolas Torquet, Thibault Lagache, Stéphane Dallongeville, Albane Imbert, Thierry Legou, Anne-Marie Le Sourd, Philippe Faure, Thomas E Bourgeron, Jean-Christophe Olivo-Marin
Preclinical studies of psychiatric disorders require the use of animal models to investigate the impact of environmental factors or genetic mutations on complex traits such as decision-making and social interactions. Here, we present a real-time method for behavior analysis of mice housed in groups that couples computer vision, machine learning and Triggered-RFID identification to track and monitor animals over several days in enriched environments. The system extracts a thorough list of individual and collective behavioral traits and provides a unique phenotypic profile for each animal. On mouse models, we study the impact of mutations of genes Shank2 and Shank3 involved in autism. Characterization and integration of data from behavioral profiles of mutated female mice reveals distinctive activity levels and involvement in complex social configuration.
7,088 downloads bioRxiv animal behavior and cognition
Researchers and educators have long wrestled with the question of how best to teach their clients be they human, animal or machine. Here we focus on the role of a single variable, the difficulty of training, and examine its effect on the rate of learning. In many situations we find that there is a sweet spot in which training is neither too easy nor too hard, and where learning progresses most quickly. We derive conditions for this sweet spot for a broad class of learning algorithms in the context of binary classification tasks, in which ambiguous stimuli must be sorted into one of two classes. For all of these gradient-descent based learning algorithms we find that the optimal error rate for training is around 15.87% or, conversely, that the optimal training accuracy is about 85%. We demonstrate the efficacy of this 'Eighty Five Percent Rule' for artificial neural networks used in AI and biologically plausible neural networks thought to describe human and animal learning.
6,760 downloads bioRxiv animal behavior and cognition
Many microbes induce striking behavioral changes in their animal hosts, but how they achieve this is poorly understood, especially at the molecular level. Mechanistic understanding has been largely constrained by the lack of a model system with advanced tools for molecular manipulation. We recently discovered a strain of the behavior-manipulating fungal pathogen Entomophthora muscae infecting wild Drosophila, and established methods to infect D. melanogaster in the lab. Lab-infected flies manifest the moribund behaviors characteristic of E. muscae infection: hours before death, they climb upward, extend their proboscides and affix in place, then raise their wings, clearing a path for infectious spores to launch from their abdomens. We found that E. muscae invades the fly nervous system, suggesting a direct means by which the fungus could induce behavioral changes. Given the vast molecular toolkit available for D. melanogaster, we believe this new system will enable rapid progress in understanding the mechanistic basis of E. muscaes behavioral manipulation in the fly.
6,439 downloads bioRxiv animal behavior and cognition
There is a growing number of dogs kept as companion animals, and the methods by which they are trained range broadly from those using mostly positive punishment and negative reinforcement (aversive-based methods) to those using primarily positive reinforcement (reward-based methods). Although the use of aversive-based methods has been strongly criticized for negatively affecting dog welfare, these claims do not find support in solid scientific evidence. Previous research on the subject lacks companion dog-focused research, investigation of the entire range of aversive-based techniques (beyond shock-collars), objective measures of welfare, and long-term welfare studies. The aim of the present study was to perform a comprehensive evaluation of the short- and long-term effects of aversive- and reward-based training methods on companion dog welfare. Ninety-two companion dogs were recruited from three reward-based (Group Reward, n=42) and four aversive-based (Group Aversive, n=50) dog training schools. For the short-term welfare assessment, dogs were video recorded for three training sessions and six saliva samples were collected, three at home (baseline levels) and three after the training sessions (post-training levels). Video recordings were then used to examine the frequency of stress-related behaviors (e.g., lip lick, yawn) and the overall behavioral state of the dog (e.g., tense, relaxed), and saliva samples were analyzed for cortisol concentration. For the long-term welfare assessment, dogs performed a cognitive bias task. Dogs from Group Aversive displayed more stress-related behaviors, spent more time in tense and low behavioral states and more time panting during the training sessions, showed higher elevations in cortisol levels after training and were more ‘pessimistic’ in the cognitive bias task than dogs from Group Reward. These findings indicate that the use of aversive-based methods compromises the welfare of companion dogs in both the short- and the long-term.
6,328 downloads bioRxiv animal behavior and cognition
The ability to perceive and recognise a reflected mirror image as self (mirror self-recognition, MSR) is considered a hallmark of cognition across species. Although MSR has been reported in mammals and birds, it is not known to occur in any other major taxon. A factor potentially limiting the ability to test for MSR is that the established assay for MSR, the mark test, shows an interpretation bias towards animals with the dexterity (or limbs) required to touch a mark. Here, we show that the cleaner wrasse fish, Labroides dimidiatus, passes through all phases of the mark test: (i) social reactions towards the reflection, (ii) repeated idiosyncratic behaviours towards the mirror (contingency testing), and (iii) frequent observation of their reflection. When subsequently provided with a coloured tag, individuals attempt to remove the mark in the presence of a mirror but show no response towards transparent marks, or to coloured marks in the absence of a mirror. This remarkable finding presents a challenge to our interpretation of the mark test – do we accept that these behavioural responses in the mark test, which are taken as evidence of self-recognition in other species, mean that fish are self-aware? Or do we conclude that these behavioural patterns have a basis in a cognitive process other than self-recognition? If the former, what does this mean for our understanding of animal intelligence? If the latter, what does this mean for our application and interpretation of the mark test as a metric for animal cognitive abilities?
5,962 downloads bioRxiv animal behavior and cognition
Pose estimation is crucial for many applications in neuroscience, biomechanics, genetics and beyond. We recently presented a highly efficient method for markerless pose estimation based on transfer learning with deep neural networks called DeepLabCut. Current experiments produce vast amounts of video data, which pose challenges for both storage and analysis. Here we improve the inference speed of DeepLabCut by up to tenfold and benchmark these updates on various CPUs and GPUs. In particular, depending on the frame size, poses can be inferred offline at up to 1200 frames per second (FPS). For instance, 278 x 278 images can be processed at 225 FPS on a GTX 1080 Ti graphics card. Furthermore, we show that DeepLabCut is highly robust to standard video compression (ffmpeg). Compression rates of greater than 1,000 only decrease accuracy by about half a pixel (for 640 x 480 frame size). DeepLabCut's speed and robustness to compression can save both time and hardware expenses.
5,572 downloads bioRxiv animal behavior and cognition
Habits form a crucial component of behavior. In recent years, key computational models have conceptualized habits as arising from model-free reinforcement learning (RL) mechanisms, which typically select between available actions based on the future value expected to result from each. Traditionally, however, habits have been understood as behaviors that can be triggered directly by a stimulus, without requiring the animal to evaluate expected outcomes. Here, we develop a computational model instantiating this traditional view, in which habits develop through the direct strengthening of recently taken actions rather than through the encoding of outcomes. We demonstrate that this model accounts for key behavioral manifestations of habits, including insensitivity to outcome devaluation and contingency degradation, as well as the effects of reinforcement schedule on the rate of habit formation. The model also explains the prevalent observation of perseveration in repeated-choice tasks as an additional behavioral manifestation of the habit system. We suggest that mapping habitual behaviors onto value-free mechanisms provides a parsimonious account of existing behavioral and neural data. This mapping may provide a new foundation for building robust and comprehensive models of the interaction of habits with other, more goal-directed types of behaviors and help to better guide research into the neural mechanisms underlying control of instrumental behavior more generally.
4,866 downloads bioRxiv animal behavior and cognition
Behavior is a unifying organismal process where genes, neural function, anatomy and environment converge and interrelate. Here we review the current state and discuss the future impact of accelerating advances in technology for behavioral studies, focusing on rodents as an exemplar. We frame our perspective in three dimensions: degree of experimental constraint, dimensionality of data, and level of description. We argue that "big behavioral data" presents challenges proportionate to its promise and describe how these challenges might be met through opportunities afforded by the two rival conceptual legacies of 20th century behavioral science, ethology and psychology. We conclude that although "more is not necessarily better", copious, quantitative and open behavioral data has the potential to transform and unify these two disciplines and to solidify the foundations of others, including neuroscience, but only if the development of novel theoretical frameworks and improved experimental designs matches the technological progress.
4,833 downloads bioRxiv animal behavior and cognition
Automated visual tracking of animals is rapidly becoming an indispensable tool for the study of behavior. It offers a quantitative methodology by which organisms' sensing and decision-making can be studied in a wide range of ecological contexts. Despite this, existing solutions tend to be challenging to deploy in practice, especially when considering long and/or high-resolution video streams. Here, we present TRex, a fast and easy-to-use solution for tracking a large number of individuals simultaneously with real-time (60Hz) tracking performance for up to approximately 256 individuals and estimates 2D body postures and visual fields, both in open and closed-loop contexts. Additionally, TRex offers highly-accurate, deep-learning-based visual identification of up to approximately 100 unmarked individuals, where it is between 2.5-46.7 times faster, and requires 2-10 times less memory, than comparable software (with relative performance increasing for more organisms and longer videos) and provides interactive data-exploration within an intuitive, platform-independent graphical user interface. ### Competing Interest Statement The authors have declared no competing interest.
4,554 downloads bioRxiv animal behavior and cognition
Benedict C Jones, Amanda C Hahn, Claire I Fisher, Hongyi Wang, Michal Kandrik, Chengyang Han, Vanessa Fasolt, Danielle Morrison, Anthony J Lee, Iris J Holzleitner, Kieran J O’Shea, S. Craig Roberts, Anthony C Little, Lisa M DeBruine
Although widely cited as strong evidence that sexual selection has shaped human facial attractiveness judgments, evidence that preferences for masculine characteristics in men's faces are related to women's hormonal status is equivocal and controversial. Consequently, we conducted the largest ever longitudinal study of the hormonal correlates of women's preferences for facial masculinity (N=584). Analyses showed no compelling evidence that preferences for facial masculinity were related to changes in women's salivary steroid hormone levels. Furthermore, both within-subject and between-subject comparisons showed no evidence that oral contraceptive use decreased masculinity preferences. However, women generally preferred masculinized over feminized versions of men's faces, particularly when assessing men's attractiveness for short-term, rather than long-term, relationships. Our results do not support the hypothesized link between women's preferences for facial masculinity and their hormonal status.
4,480 downloads bioRxiv animal behavior and cognition
Scientific datasets are growing rapidly in scale and complexity. Consequently, the task of understanding these data to answer scientific questions increasingly requires the use of compression algorithms that reduce dimensionality by combining correlated features and cluster similar observations to summarize large datasets. Here we introduce a method for both dimension reduction and clustering called VAE-SNE (variational autoencoder stochastic neighbor embedding). Our model combines elements from deep learning, probabilistic inference, and manifold learning to produce interpretable compressed representations while also readily scaling to tens-of-millions of observations. Unlike existing methods, VAE-SNE simultaneously compresses high-dimensional data and automatically learns a distribution of clusters within the data \---| without the need to manually select the number of clusters. This naturally creates a multi-scale representation, which makes it straightforward to generate coarse-grained descriptions for large subsets of related observations and select specific regions of interest for further analysis. VAE-SNE can also quickly and easily embed new samples, detect outliers, and can be optimized with small batches of data, which makes it possible to compress datasets that are otherwise too large to fit into memory. We evaluate VAE-SNE as a general purpose method for dimensionality reduction by applying it to multiple real-world datasets and by comparing its performance with existing methods for dimensionality reduction. We find that VAE-SNE produces high-quality compressed representations with results that are on par with existing nonlinear dimensionality reduction algorithms. As a practical example, we demonstrate how the cluster distribution learned by VAE-SNE can be used for unsupervised action recognition to detect and classify repeated motifs of stereotyped behavior in high-dimensional timeseries data. Finally, we also introduce variants of VAE-SNE for embedding data in polar (spherical) coordinates and for embedding image data from raw pixels. VAE-SNE is a robust, feature-rich, and scalable method with broad applicability to a range of datasets in the life sciences and beyond. ### Competing Interest Statement The authors have declared no competing interest.
4,479 downloads bioRxiv animal behavior and cognition
Aberrant social behavior is a core feature of many neuropsychiatric disorders, yet the study of complex social behavior in freely moving rodents is relatively infrequently incorporated into preclinical models. This likely contributes to limited translational impact. A major bottleneck for the adoption of socially complex, ethology-rich, preclinical procedures are the technical limitations for consistently annotating detailed behavioral repertoires of rodent social behavior. Manual annotation is subjective, prone to observer drift, and extremely time-intensive. Commercial approaches are expensive and inferior to manual annotation. Open-source alternatives often require significant investments in specialized hardware and significant computational and programming knowledge. By combining recent computational advances in convolutional neural networks and pose-estimation with further machine learning analysis, complex rodent social behavior is primed for inclusion under the umbrella of computational neuroethology. Here we present an open-source package with graphical interface and workflow (Simple Behavioral Analysis, SimBA) that uses pose-estimation to create supervised machine learning predictive classifiers of rodent social behavior, with millisecond resolution and accuracies that can out-perform human observers. SimBA does not require specialized video acquisition hardware nor extensive computational background. Standard descriptive statistical analysis, along with graphical region of interest annotation, are provided in addition to predictive classifier generation. To increase ease-of-use for behavioural neuroscientists, we designed SimBA with accessible menus for pre-processing videos, annotating behavioural training datasets, selecting advanced machine learning options, robust classifier validation functions and flexible visualizations tools. This allows for predictive classifier transparency, explainability and tunability prior to, and during, experimental use. We demonstrate that this approach is flexible and robust in both mice and rats by classifying social behaviors that are commonly central to the study of brain function and social motivation. Finally, we provide a library of pose-estimation weights and behavioral predictive classifiers for resident-intruder behaviors in mice and rats. All code and data, together with detailed tutorials and documentation, are available on the SimBA GitHub repository. ### Competing Interest Statement The authors have declared no competing interest.
4,337 downloads bioRxiv animal behavior and cognition
Quantitative behavioral measurements are important for answering questions across scientific disciplines—from neuroscience to ecology. State-of-the-art deep-learning methods offer major advances in data quality and detail by allowing researchers to automatically estimate locations of an animal’s body parts directly from images or videos. However, currently-available animal pose estimation methods have limitations in speed and robustness. Here we introduce a new easy-to-use software toolkit, DeepPoseKit , that addresses these problems using an eZcient multi-scale deep-learning model, called Stacked DenseNet , and a fast GPU-based peak-detection algorithm for estimating keypoint locations with subpixel precision. These advances improve processing speed >2× with no loss in accuracy compared to currently-available methods. We demonstrate the versatility of our methods with multiple challenging animal pose estimation tasks in laboratory and field settings—including groups of interacting individuals. Our work reduces barriers to using advanced tools for measuring behavior and has broad applicability across the behavioral sciences.
4,303 downloads bioRxiv animal behavior and cognition
Absolute pitch (AP) refers to the rare ability to name the pitch of a tone without external reference. It is widely believed that acquiring AP in adulthood is impossible, since AP is only for the selected few with rare genetic makeup and early musical training. In three experiments, we trained adults to name pitches for 12 to 40 hours. Within the training period, 14% of the participants were able to name twelve pitches at 90% accuracy or above, a performance level comparable with typical AP possessors. At the group level, performance enhancement, learning generalization and sustainability were observed as in typical perceptual learning studies. The findings suggest that AP continues to be learnable in adulthood. The genesis of AP may be better explained by the amount and type of perceptual experience.
4,209 downloads bioRxiv animal behavior and cognition
Dogs are hypersocial with humans, and their integration into human social ecology makes dogs a unique model for studying cross-species social bonding. However, the proximal neural mechanisms driving dog-human social interaction are unknown. We used fMRI in 15 awake dogs to probe the neural basis for their preferences for social interaction and food reward. In a first experiment, we used the ventral caudate as a measure of intrinsic reward value and compared activation to conditioned stimuli that predicted food, praise, or nothing. Relative to the control stimulus, the caudate was significantly more active to the reward-predicting stimuli and showed roughly equal or greater activation to praise versus food in 13 of 15 dogs. To confirm that these differences were driven by the intrinsic value of social praise, we performed a second imaging experiment in which the praise was withheld on a subset of trials. The difference in caudate activation to the receipt of praise, relative to its withholding, was strongly correlated with the differential activation to the conditioned stimuli in the first experiment. In a third experiment, we performed an out-of-scanner choice task in which the dog repeatedly selected food or owner in a Y-maze. The relative caudate activation to food- and praise-predicting stimuli in Experiment 1 was a strong predictor of each dog's sequence of choices in the Y-maze. Analogous to similar neuroimaging studies of individual differences in human social reward, our findings demonstrate a neural mechanism for preference in domestic dogs that is stable within, but variable between, individuals. Moreover, the individual differences in the caudate responses indicate the potentially higher value of social than food reward for some dogs and may help to explain the apparent efficacy of social interaction in dog training.
4,128 downloads bioRxiv animal behavior and cognition
Animals produce vocalizations that range in complexity from a single repeated call to hundreds of unique vocal elements patterned in sequences unfolding over hours. Characterizing complex vocalizations can require considerable effort and a deep intuition about each species’ vocal behavior. Even with a great deal of experience, human characterizations of animal communication can be affected by human perceptual biases. We present here a set of computational methods that center around projecting animal vocalizations into low dimensional latent representational spaces that are directly learned from data. We apply these methods to diverse datasets from over 20 species, including humans, bats, songbirds, mice, cetaceans, and nonhuman primates, enabling high-powered comparative analyses of unbiased acoustic features in the communicative repertoires across species. Latent projections uncover complex features of data in visually intuitive and quantifiable ways. We introduce methods for analyzing vocalizations as both discrete sequences and as continuous latent variables. Each method can be used to disentangle complex spectro-temporal structure and observe long-timescale organization in communication. Finally, we show how systematic sampling from latent representational spaces of vocalizations enables comprehensive investigations of perceptual and neural representations of complex and ecologically relevant acoustic feature spaces.
- 27 Nov 2020: The website and API now include results pulled from medRxiv as well as bioRxiv.
- 18 Dec 2019: We're pleased to announce PanLingua, a new tool that enables you to search for machine-translated bioRxiv preprints using more than 100 different languages.
- 21 May 2019: PLOS Biology has published a community page about Rxivist.org and its design.
- 10 May 2019: The paper analyzing the Rxivist dataset has been published at eLife.
- 1 Mar 2019: We now have summary statistics about bioRxiv downloads and submissions.
- 8 Feb 2019: Data from Altmetric is now available on the Rxivist details page for every preprint. Look for the "donut" under the download metrics.
- 30 Jan 2019: preLights has featured the Rxivist preprint and written about our findings.
- 22 Jan 2019: Nature just published an article about Rxivist and our data.
- 13 Jan 2019: The Rxivist preprint is live!