Automated procedure to assess pup retrieval in laboratory mice

Simple Behavioral Analysis (SimBA) as a platform for explainable machine learning in behavioral neuroscience

The study of complex behaviors is often challenging when using manual annotation due to the absence of quantifiable behavioral definitions and the subjective nature of behavioral annotation. Integration of supervised machine learning approaches mitigates some of these issues through the inclusion of accessible and explainable model interpretation. To decrease barriers to access, and with an emphasis on accessible model explainability, we developed the open-source Simple Behavioral Analysis (SimBA) platform for behavioral neuroscientists. SimBA introduces several machine learning interpretability tools, including SHapley Additive exPlanation (SHAP) scores, that aid in creating explainable and transparent behavioral classifiers. Here we show how the addition of explainability metrics allows for quantifiable comparisons of aggressive social behavior across research groups and species, reconceptualizing behavior as a sharable reagent and providing an open-source framework. We provide an open-source, graphical user interface (GUI)-driven, well-documented package to facilitate the movement toward improved automation and sharing of behavioral classification tools across laboratories.

Effects of a TAML catalyst on mice exposed during pregnancy and lactation

Tetra-amido macrocyclic ligands (TAMLs) are catalysts designed to mimic endogenous peroxidases that can degrade pollutants. Before TAMLs gain widespread use, it is first important to determine if they have endocrine disrupting properties. In this study, we evaluated the effects of the iron TAML, NT7, on hormone-sensitive outcomes in mice exposed during pregnancy and lactation, and on their litters prior to weaning. We administered NT7 at one of three doses to mice via drinking water prior to and then throughout pregnancy and lactation. Two hormonally active pharmaceuticals, ethinyl estradiol (EE2) and flutamide (FLUT), a known estrogen receptor agonist and androgen receptor antagonist, respectively, were also included. In the females, we measured pre- and post-parturition weight, length of pregnancy, organ weights at necropsy, and morphology of the mammary gland at the end of the lactational period. We also quantified maternal behaviors at three stages of lactation. For the offspring, we measured litter size, litter weights, and the achievement of other developmental milestones. We observed only one statistically significant effect of NT7, a decrease in the percentage of pups with ear opening at postnatal day 5. This contrasts with the numerous effects of EE2 on both the mother and the litter, as well as several modest effects of FLUT. The approach taken in this study could provide guidance for future studies that aim to evaluate novel compounds for endocrine disrupting properties.

Multidimensional Analysis of a Social Behavior Identifies Regression and Phenotypic Heterogeneity in a Female Mouse Model for Rett Syndrome

Regression is a key feature of neurodevelopmental disorders such as autism spectrum disorder, Fragile X syndrome, and Rett syndrome (RTT). RTT is caused by mutations in the X-linked gene methyl-CpG-binding protein 2 (MECP2). It is characterized by an early period of typical development with subsequent regression of previously acquired motor and speech skills in girls. The syndromic phenotypes are individualistic and dynamic over time. Thus far, it has been difficult to capture these dynamics and syndromic heterogeneity in the preclinical Mecp2-heterozygous female mouse model (Het). The emergence of computational neuroethology tools allows for robust analysis of complex and dynamic behaviors to model endophenotypes in preclinical models. Toward this first step, we utilized DeepLabCut, a marker-less pose estimation software to quantify trajectory kinematics and multidimensional analysis to characterize behavioral heterogeneity in Het in the previously benchmarked, ethologically relevant social cognition task of pup retrieval. We report the identification of two distinct phenotypes of adult Het: Het that display a delay in efficiency in early days and then improve over days like wild-type mice and Het that regress and perform worse in later days. Furthermore, regression is dependent on age and behavioral context and can be detected in the initial days of retrieval. Together, the novel identification of two populations of Het suggests differential effects on neural circuitry, opens new avenues to investigate the underlying molecular and cellular mechanisms of heterogeneity, and designs better studies for stratifying therapeutics.

High-THC Cannabis Smoke Impairs Incidental Memory Capacity in Spontaneous Tests of Novelty Preference for Objects and Odors in Male Rats

Working memory is an executive function that orchestrates the use of limited amounts of information, referred to as working memory capacity, in cognitive functions. Cannabis exposure impairs working memory in humans; however, it is unclear whether Cannabis facilitates or impairs rodent working memory and working memory capacity. The conflicting literature in rodent models may be at least partly because of the use of drug exposure paradigms that do not closely mirror patterns of human Cannabis use. Here, we used an incidental memory capacity paradigm where a novelty preference is assessed after a short delay in spontaneous recognition-based tests. Either object or odor-based stimuli were used in test variations with sets of identical [identical stimuli test (IST)] and different [different stimuli test (DST)] stimuli (three or six) for low-memory and high-memory loads, respectively. Additionally, we developed a human-machine hybrid behavioral quantification approach which supplements stopwatch-based scoring with supervised machine learning-based classification. After validating the spontaneous IST and DST in male rats, 6-item test versions with the hybrid quantification method were used to evaluate the impact of acute exposure to high-Δ9-tetrahydrocannabinol (THC) or high-CBD Cannabis smoke on novelty preference. Under control conditions, male rats showed novelty preference in all test variations. We found that high-THC, but not high-CBD, Cannabis smoke exposure impaired novelty preference for objects under a high-memory load. Odor-based recognition deficits were seen under both low-memory and high-memory loads only following high-THC smoke exposure. Ultimately, these data show that Cannabis smoke exposure impacts incidental memory capacity of male rats in a memory load-dependent, and stimuli-specific manner.

Automated behavioral scoring: Do we even need humans?

The development of automated behavior scoring technology has been a tremendous boon to the study of social behavior. However, completely outsourcing behavioral analysis to a computer runs the risk of overlooking important nuances, and researchers risk distancing themselves from their very object of study. Here, I make the case that while automating analysis has been valuable, and overautomating analysis is risky, more effort should be spent automating the collection of behavioral data. Continuous automated behavioral observations conducted in situ have the promise to reduce confounding elements of social behavior research, such as handling stress, novel environments, one-time "snapshot" measures, and experimenter presence. Now that we have the capability to automatically process behavioral observations thanks to machine vision and machine learning, we would do well to leverage the same open-source ethos to increase the throughput of behavioral observation and collection. Fortunately, several such platforms have recently been developed. Repeated testing in the home environment will produce higher qualities and quantities of data, bringing us closer to realizing the ethological goals of studying animal behavior in a naturalistic context.

PiE: an open-source pipeline for home cage behavioral analysis

Over the last two decades a growing number of neuroscience labs are conducting behavioral assays in rodents. The equipment used to collect this behavioral data must effectively limit environmental and experimenter disruptions, to avoid confounding behavior data. Proprietary behavior boxes are expensive, offer limited compatible sensors, and constrain analysis with closed-source hardware and software. Here, we introduce PiE, an open-source, end-to-end, user-configurable, scalable, and inexpensive behavior assay system. The PiE system includes the custom-built behavior box to hold a home cage, as well as software enabling continuous video recording and individual behavior box environmental control. To limit experimental disruptions, the PiE system allows the control and monitoring of all aspects of a behavioral experiment using a remote web browser, including real-time video feeds. To allow experiments to scale up, the PiE system provides a web interface where any number of boxes can be controlled, and video data easily synchronized to a remote location. For the scoring of behavior video data, the PiE system includes a standalone desktop application that streamlines the blinded manual scoring of large datasets with a focus on quality control and assay flexibility. The PiE system is ideal for all types of behavior assays in which video is recorded. Users are free to use individual components of this setup independently, or to use the entire pipeline from data collection to analysis. Alpha testers have included scientists without prior coding experience. An example pipeline is demonstrated with the PiE system enabling the user to record home cage maternal behavior assays, synchronize the resulting data, conduct blinded scoring, and import the data into R for data visualization and analysis.

A Layered, Hybrid Machine Learning Analytic Workflow for Mouse Risk Assessment Behavior

Accurate and efficient quantification of animal behavior facilitates the understanding of the brain. An emerging approach within machine learning (ML) field is to combine multiple ML-based algorithms to quantify animal behavior. These so-called hybrid models have emerged because of limitations associated with supervised [e.g., random forest (RF)] and unsupervised [e.g., hidden Markov model (HMM)] ML models. For example, RF models lack temporal information across video frames, and HMM latent states are often difficult to interpret. We sought to develop a hybrid model, and did so in the context of a study of mouse risk assessment behavior. We used DeepLabCut to estimate the positions of mouse body parts. Positional features were calculated using DeepLabCut outputs and were used to train RF and HMM models with equal number of states, separately. The per-frame predictions from RF and HMM models were then passed to a second HMM model layer ("reHMM"). The outputs of the reHMM layer showed improved interpretability over the initial HMM output. Finally, we combined predictions from RF and HMM models with selected positional features to train a third HMM model ("reHMM+"). This reHMM+ layered hybrid model unveiled distinctive temporal and human-interpretable behavioral patterns. We applied this workflow to investigate risk assessment to trimethylthiazoline and snake feces odor, finding unique behavioral patterns to each that were separable from attractive and neutral stimuli. We conclude that this layered, hybrid ML workflow represents a balanced approach for improving the depth and reliability of ML classifiers in chemosensory and other behavioral contexts.

BAMBI: A new method for automated assessment of bidirectional early-life interaction between maternal behavior and pup vocalization in mouse dam-pup dyads

Vital early-life dyadic interaction in mice requires a pup to signal its needs adequately, and a dam to recognize and respond to the pup's cues accurately and timely. Previous research might have missed important biological and/or environmental elements of this complex bidirectional interaction, because it often focused on one dyadic member only. In laboratory rodents, the Pup Retrieval Test (PRT) is the leading procedure to assess pup-directed maternal care. The present study describes BAMBI (Bidirectional Automated Mother-pup Behavioral Interaction test), a novel automated PRT methodology based on synchronous video recording of maternal behavior and audio recording of pup vocalizations, which allows to assess bidirectional dam-pup dyadic interaction. We were able to estimate pup retrieval and pup vocalization parameters accurately in 156 pups from 29 dams on postnatal days (PND) 5, 7, 9, 11, and 13. Moreover, we showed an association between number of emitted USVs and retrieval success, indicating dyadic interdependency and bidirectionality. BAMBI is a promising new automated home-cage behavioral method that can be applied to both basic and preclinical studies investigating complex phenotypes related to early-life social development.

Using deep learning to study emotional behavior in rodent models

Quantifying emotional aspects of animal behavior (e.g., anxiety, social interactions, reward, and stress responses) is a major focus of neuroscience research. Because manual scoring of emotion-related behaviors is time-consuming and subjective, classical methods rely on easily quantified measures such as lever pressing or time spent in different zones of an apparatus (e.g., open vs. closed arms of an elevated plus maze). Recent advancements have made it easier to extract pose information from videos, and multiple approaches for extracting nuanced information about behavioral states from pose estimation data have been proposed. These include supervised, unsupervised, and self-supervised approaches, employing a variety of different model types. Representations of behavioral states derived from these methods can be correlated with recordings of neural activity to increase the scope of connections that can be drawn between the brain and behavior. In this mini review, we will discuss how deep learning techniques can be used in behavioral experiments and how different model architectures and training paradigms influence the type of representation that can be obtained.

Estimating genetics of body dimensions and activity levels in pigs using automated pose estimation

Pig breeding is changing rapidly due to technological progress and socio-ecological factors. New precision livestock farming technologies such as computer vision systems are crucial for automated phenotyping on a large scale for novel traits, as pigs’ robustness and behavior are gaining importance in breeding goals. However, individual identification, data processing and the availability of adequate (open source) software currently pose the main hurdles. The overall goal of this study was to expand pig weighing with automated measurements of body dimensions and activity levels using an automated video-analytic system: DeepLabCut. Furthermore, these data were coupled with pedigree information to estimate genetic parameters for breeding programs. We analyzed 7428 recordings over the fattening period of 1556 finishing pigs (Piétrain sire x crossbred dam) with two-week intervals between recordings on the same pig. We were able to accurately estimate relevant body parts with an average tracking error of 3.3 cm. Body metrics extracted from video images were highly heritable (61–74%) and significantly genetically correlated with average daily gain (r_g = 0.81–0.92). Activity traits were low to moderately heritable (22–35%) and showed low genetic correlations with production traits and physical abnormalities. We demonstrated a simple and cost-efficient method to extract body dimension parameters and activity traits. These traits were estimated to be heritable, and hence, can be selected on. These findings are valuable for (pig) breeding organizations, as they offer a method to automatically phenotype new production and behavioral traits on an individual level.

Toward the explainability, transparency, and universality of machine learning for behavioral classification in neuroscience

The use of rigorous ethological observation via machine learning techniques to understand brain function (computational neuroethology) is a rapidly growing approach that is poised to significantly change how behavioral neuroscience is commonly performed. With the development of open-source platforms for automated tracking and behavioral recognition, these approaches are now accessible to a wide array of neuroscientists despite variations in budget and computational experience. Importantly, this adoption has moved the field toward a common understanding of behavior and brain function through the removal of manual bias and the identification of previously unknown behavioral repertoires. Although less apparent, another consequence of this movement is the introduction of analytical tools that increase the explainabilty, transparency, and universality of the machine-based behavioral classifications both within and between research groups. Here, we focus on three main applications of such machine model explainabilty tools and metrics in the drive toward behavioral (i) standardization, (ii) specialization, and (iii) explainability. We provide a perspective on the use of explainability tools in computational neuroethology, and detail why this is a necessary next step in the expansion of the field. Specifically, as a possible solution in behavioral neuroscience, we propose the use of Shapley values via Shapley Additive Explanations (SHAP) as a diagnostic resource toward explainability of human annotation, as well as supervised and unsupervised behavioral machine learning analysis.