Do behavioral test scores represent repeatable phenotypes of female mice?

The marble bury test is a commonly applied behavioral test often used to screen pharmaceuticals for treatment of compulsivity or anxiety disorders and to better understand the underlying neurobiology of these conditions using rodent models. We explore the use of the marble bury test in a repeated fashion over longer time intervals to assess how it may be used in a more chronic context. Assuming that marble bury scores represent a neurobiological phenotype of an individual, we hypothesize that the measurement of this phenotype is repeatable over time in a healthy animal. We performed four trials over the course of 12 weeks, using three overlapping scores for marble burying: 100%, >50%, and >0% buried. Despite intertrial differences in the mean number of marbles buried, we found significant repeatability scores across 12 weeks for two measurements: marbles buried 100% and marbles buried >50%. Exploration of pairwise repeatabilities between the trials revealed highest repeatability during shorter time intervals. Although the effect of time is difficult to disentangle from possible effects of body size/development on the scores, we hypothesize that a combination of these two factors together influence repeatability of scores and should be explored further. Understanding the patterns in repeatability in marble burying scores can inform experimental design to use this test over longer time intervals for chronic conditions or long-term interventions. Additionally, these data indicate that a pre-test, post-test design is possible, which can reduce the number of animals needed for research applications while maintaining statistical power.