The lack of appropriate animal models for bipolar disorder (BPD) hinders the translation of novel molecular and genetic findings into the development of new more efficient treatments. Attempts to develop a comprehensive model for BPD did not result in a practical and valid model and at present most studies utilize a limited number of models for specific components of the disorder.

Whereas there is a higher availability of models for the depression pole of BPD, only a few models represent the manic pole with the most frequently used being psychostimulant-induced hyperactivity. This last model had been important in studies of the disease and has some validity but it is clear that by itself cannot be considered to represent mania. Additional models for facets of BPD are needed to allow better screening of new drugs and new mutant mice. Such models may also support the exploration of endophenotypes of BPD and the mechanisms of the disease. An advantage of a battery approach is that each model can be only partially valid when used alone but the combination of a few models may result in strong validity.

The present study suggests that such a battery can be based on existing models previously developed in the context of studying normal behavior or other disorders after an initial validation in the context of BPD. An example for this idea is described using the resident–intruder test for aggression. Present results show that 3 weeks oral treatment with 1.2–2.4% lithium (increasing doses), or 20 g/kg daily dose of valproate, significantly reduced aggressive behavior in resident mice without affecting non-aggressive social interactions. Accordingly, it is suggested that the simplified resident–intruder paradigm may model the aggression related to mania as part of a test battery for facets of BPD. It is further speculated that, pending further research, this paradigm can be combined with additional methods to explore changes in the LHPA axis that may be linked to an important endophenotype of BPD.