These lines of evidence suggested that the Tg mice fulfill the three validity criteria. With regard to structural validity, the mice are at least appropriate as a model of mood symptoms associated with hCPEO. Detection of mtDNA deletions in the brains of patients with bipolar disorder (Kato et al., 1997) further support this validity. However, construct validity may be compromised by the fact that we have not yet found a bipolar disorder patient carrying the D198A mutation of POLG. The mice also at least partly fulfill the face validity criterion, because the periodic activity change looks like bipolar disorder, and the antidepressant-induced manic switch-like phenomenon is a hallmark of bipolar disorder. Circadian rhythm abnormality is also reported in bipolar disorder. The predictive validity is satisfied by the effect of lithium on circadian rhythm and periodic activity change. These findings together suggest that the Tg mice have potential as bipolar disorder model mice (Kasahara et al., 2006).