"Personality" in laboratory mice used for biomedical research: a way of understanding variability?

The mouse, including countless lines of transgenic and knockout mice, has become the most prominent model organism in biomedical research. Behavioral characterization is often conducted in batteries of short tests on locomotion, anxiety, learning and memory, etc. In such tests, any individual differences within groups are usually considered to be disturbing variance. In order to reduce variance in experimental animal research enormous efforts of standardization have been made. While a substantial reduction of variability has been reached compared to the earlier years of experimental animal studies a considerable amount of inter-individual differences still seems to escape standardization. This effect is demonstrated and evaluated by re-analyzing data from two experiments conducted in our laboratory with inbred mice. Interestingly, behavioral patterns of individual animals seem to be correlated across context and time. In evolutionary biology, "animal personalities" have been discussed recently to comprise such stable patterns. We argue here, that nonrandom behavioral correlations across contexts and time might underlie the variability commonly found in biomedical mouse studies.