Rodent Behavioral Core: The Analysis of Behavior

National Institute Of Mental Health

Investigators

Linked publications, trials & patents

Abstract

Animals evolved in complex environments, producing a wide range of behaviors, including navigation, foraging, prey capture, and conspecific interactions, which vary over timescales ranging from milliseconds to days. Historically, these behaviors have been the focus of study for ecology and ethology, while systems neuroscience has largely focused on short timescale behaviors that can be repeated thousands of times and occur in highly artificial environments. Recent advances in machine learning, miniaturized implants, and motor behavior tracking, have made it possible to study freely moving animals in more natural conditions while applying systems techniques to enable temporally specific perturbations, modeling behavioral strategies, and recording from large numbers of neurons while animals are freely moving. Combined with genetically tractable rodent models of neurodevelopmental, neuropsychiatric, and neurodegenerative diseases, these techniques bring a new level of understanding of how behavioral neuroscience links systems-level circuitry to behavior, cognition and emotion. Complex cognitive behavior in rodents is often gauged by measuring the pattern of behavioral responses in tasks that involve, for example, decision-making, attention, memory, rule learning, flexibility, discrimination, and problem solving. In these tasks, rats and mice typically indicate their decisions by nose-poking visual patterns on a touchscreen like an iPad, making nose-poke entries into a series of lit holes, or depressing an extended lever triggered by time or cues. Some cognitive functions extrapolated from animal behavior have positively informed our investigation of cognitive functions in humans. Such animal-to-human approaches (e.g., delayed response) have directed the design and development of analogous tests for use in humans (e.g., self-ordered working memory). Behavioral neuroscience has also benefitted in the opposite direction by means of human-to-animal approaches as in the case of extradimensionsal/intradimensional set shifting, a test based upon the principles of the human Wisconsin Card Sorting Task. Together, these advances in behavioral testing have been particularly useful in establishing the neuroanatomical and neurochemical pathology for specific cognitive deficits in a range of brain and behavior disorders. In addition to providing equipment resources, we write custom code for several users to enable detailed levels of behavioral quantification or analyses for their experiments. For example, general motor activity data is usually indexed as duration and location of activity. With custom code, we’ve been able to provide researchers with additional measures such as speed of activity as well as visual patterns of movement. For some users, we have integrated observed behavior with deep learning methods using an open-source computer vision technique to accurately quantify behavior using a pose estimation of body parts. For those measuring calcium signals in freely moving animals, we are creating a pipeline suitable for a wide range of photometry approaches to address the challenges associated with analyzing and investigating large data sets, without the need for custom analysis. We continue to design and develop cutting edge behavioral methods and applications while maintaining facility resources at a high level of utility for users at all levels of expertise. Our most recent technical advance is an automated three-chamber looming task integrated with risk that is not commercially available. It’s pneumatically controlled sliding doors open and close the safe zone and the food delivery zone which allow for dynamic control of the doors in response to real time behaviors. It also comprises a stepper motor connected to a tether to allow seamless tracking of the animal movements without introducing any friction. A total of 18 sensors and 12 controllable components allows users the flexibility to tailor the system to their specific needs, and six carefully positioned cameras enable methods of pose estimation to track and model animal movements using deep learning. Introducing new methods and technology to our behavioral repertoire requires constant maintenance and calibration of equipment, user education and interaction, and a commitment to setting the standard as the best Rodent Behavioral Core facility in the world in terms of research quality. In the past year, we have supported labs of several principal Investigators from NIMH, as well as NINDS, NIA, NIDCD, NHGRI, NICHD, NIDCD, NEI, NHLBI, NIAID, NCI, NIBIB and NIDDK to further their scientific goals. Over the past year, we custom designed and developed new tasks to measure various aspects of behavior including providing options for users to assess certain emotional behaviors such as defensive reactions that were more ethologically grounded. More recently, these tests have been redesigned to allow frame-by-frame analysis to provide a more refined behavioral output. Virtually every piece of equipment in the RBC including operant chambers, mazes and open testing arenas can be integrated with transistor-transistor logic (TTL) capability to interface with any TTL triggerable piece of equipment or software to enable optogenetic capability. We have also supported researchers in their quest to measure cellular dynamics by integrating methods of fiber photometry calcium signaling with spatial mazes, operant chambers, and tests of emotional memory to enable scientists to detect changes of fluctuations in fluorescence in intracellular calcium, which serves an indirect indicator of neural activity. The capability of optogenetic and fiber photometry studies in the RBC has made it an invaluable resource for many users who do not have the equipment or expertise to establish the infrastructure in their own labs. We have taken steps to ensure the validity and reliability of experiments conducted in the facility. First, we have improved several apparatuses to improve tracking and reduce experimental error. Second, we improved the lighting in each room to allow our users to set customized (but reproducible) lighting conditions to suit their experimental needs. This includes the addition of infrared (IR) illuminators and cameras to allow experimenters to observe animals in complete darkness. We have also expanded our ability to provide sophisticated pose estimation methods that use deep learning models. This is an open-source code computer vision technique to accurately track an animal's location as well as individual body parts. We are currently working on expanding this approach by integrating 3-dimensional imaging to capture stereoscopic effects (i.e., depth perception). This type of analysis allows us to detect and track animal behavior in 3-dimensional space. Our automated behavioral testing systems provide rich data that requires a great deal of time and effort to analyze and interpret. In the past year, we have created data processing pipelines for several of these systems such that users can easily obtain usable statistics and figures using code written by RBC staff. Its major advantage is that it allows the user to go directly from the data set creation to automatic behavioral analysis. It also provides a means to standardize behavioral testing in an open-access manner so that data generated in the RBC can be shared between collaborators. One important initiative which has potential for scientific insight relates to the automated assessment of unforced behavior in the home cage over long time intervals. it is becoming increasing apparent that assessments of cognitive and emotional behavior are most reliable when animals are in their home cage social setting. We have started on this path with multiple automated operant chambers that leverage machine-learning technology to drive the assessment of behaviors in rodents over days and weeks. Without human bias and interference, this approach reduces manual errors and increases reproducibility of with reliability while accelerating the pace of research.