Behavioral Phenotyping of Murine Disease Models with the Integrated Behavioral Station (INBEST)

The detection of functional effects in animals largely depends on the researcher’s ability to limit the variability inherent to behavioral studies. Therefore, it is important to meticulously control and minimize potential confounds that may reduce the reliability and reproducibility of behavioral data. At the same time, it is important to recognize that no test reflects a single domain of behavior, that knowledge of neurological function is mandatory, and that behavior is highly sensitive to external stressors. If the above postulates are appreciated, one may conclude that comprehensive behavioral analysis should include the time course of the response measured, as well as, involve basic functional traits and paradigms that tap into specific behavioral aspects. Many of these criteria can be fulfilled using computerized assessment of movements and behavioral acts in an enriched home-cage environment.

Thus far, it has been emphasized that behavioral phenotyping of murine models of human disease warrants additional considerations. This notion is based on the premise that functional homeostasis is challenged by internal and external stressors during disease onset. Although all potential confounds may not be eliminated by the introduction of automated, home-cage phenotyping, issues relating to inconsistent environmental settings, transportation stress, and repeated handling are minimized. This significantly enhances consistency and precision across studies; even small reductions in variability can improve detection of the effects induced by an incipient disease. Indeed, INBEST provides a wealth of information, which permits more accurate assessment of the onset, kinetics, and severity of behavioral changes, as well as important relationships among various disease-induced behavioral deficits. Reliable video-tracking depends on two lighting conditions. First, diffuse light is required in the testing room to prevent artifacts from nearby reflecting objects. Second, high color contrast can be achieved by choosing an appropriate floor color to differ from the subject color as much as possible. In our lab, this is achieved by using flood-lights positioned below the INBEST boxes and black floor trays when monitoring albino mice (white or gray background would be adequate if testing pigmented strains). With respect to the event recording aspect of INBEST, the current hardware setting (1 Piccolo video card with 4 inputs) limits 4 boxes to be concurrently used per PC. This is a rather small number of boxes, while a more suitable set-up would require 8 or even 16 cages, and thus 2 or 4 PCs, respectively. Preferably, INBEST can be used continuously over 24-hr as the home-cage. This would allow animals to habituate fully to the environment and establish stable, circadian behavioral patterns, which can be analyzed in an unbiased manner. To prevent computer data loss due to power outage, a continuous power supply (or at least an uninterruptible power source) must be secured. Lastly, to ensure proper assessment of daily food intake, it should be noted that the size of food pellets should not exceed the size of the holes in the food dispenser (the recommended size of a single food pellet is 20 mg).

It is not to be overlooked, however, that this analysis should also incorporate how different INBEST measures may interact with each other. For example, mice that spend more time in the running wheel are likely to ingest higher quantities of food and water to meet their increased caloric demands. Similarly, mice ingesting more sucrose solution may decrease their intake of food. The interpretation of these results may further be complicated by the general improvement of performance over time, particularly in regards to ingestive behavior and running wheel activity. Given their incentive-properties, experimenters may also consider limiting access to sucrose solution and the running wheel to counteract the likelihood of post-ingestive effects and excessive weight loss, respectively. However, these concerns may be more relevant in some strains than others because different strains of mice have dissimilar behavioral profiles. Although performing both baseline and experimental assessment controls for many of the issues above, experimenters need to recognize that these variables need to be taken into account when interpreting INBEST data. At the same time, some aspects of behavior cannot be studied in the home-cage environment, thus necessitating a combination with standard tests to complete the behavioral profile of subjects.

Computerized monitoring within standardized, but flexible environments appears to be the logical next step in contemporary behavioral analysis. Such a non-invasive, ethologically-based approach will allow researchers to observe the full repertoire of behavioral responses over a prolonged period of time. Theoretically, this can be accomplished by studying behavior in a “virtual”, enriched environment that closely resembles a natural habitat. Several research groups have described vision-based tracking tools that support behavioral phenotyping of mice in their home-cage ^22-25, in dyads ^{26, 27}, or in the context of large social groups ²⁸. High precision and spatial resolution can be achieved by integrating video-tracking with microchip technology for simultaneous and synchronized collection of behavioral data in a group of mice ²⁸. Thermographic cameras capable of detecting heat-signatures can be combined with implantable microchips or transponders to provide the relative location and basic physiological functions of each mouse (e.g., body temperature, heart/breathing rate). In addition, an advanced 3D tracking system would produce more accurate and quantitative recognition of behavioral acts. In order to repeatedly run a variety of tests, such a system should be automated, remotely-controlled, and modular. For example, spatial memory can be studied in larger environments by programming the appearance of distal cues on LCD walls, or by presenting/hiding dispensers with palatable food from movable floors. In a similar manner, novel objects could be presented/hidden at specific times throughout the study. Such computerized phenotyping may help in elucidating the genetic determinants of behavior, pathogenic mechanisms underlying disease models, and the development of novel therapeutic strategies. If a consensus is reached with respect to testing conditions, sequence of tests, as well as the hardware and software used, one may expect that long-awaited standardization would improve reproducibility of behavioral studies and elevate experimental psychometrics to a new level.