Trait and state anxiety in animal models: Is there correlation?

It is believed that subjects with high trait anxiety levels tend to present state anxiety reactions with greater intensity than individuals with low trait anxiety levels. In order to verify if this premise is valid for animal models of anxiety, the present work investigated the possible correlation between two behavioral tests: the elevated plus-maze, a classic model of state-anxiety, and the free-exploratory paradigm, which has been proposed as a model of trait anxiety. The behavior of 46 drug-naive, adult, Wistar, male rats was measured in these two models on two occasions, 1 week apart. Subsequently, the intraclass correlation coefficient (ICC) was calculated for the parameters "percentage of time in the novel side" (%TNS; free-exploratory paradigm), "percentage of time in the open arms" (%TOA; elevated plus-maze) and "percentage of entries into the open arms" (%EOA; elevated plus-maze). These parameters were also used to classify the animals into groups presenting high, medium or low levels of anxiety in both tests, so that the concordance between the models could be evaluated through the kappa test. The analysis resulted in low ICC (%TNSx%TOA: -0.127; %TNSx%EOA: 0.040) and low kappa index (%TNSx%TOA: -0.017; %TNSx%EOA: -0.044), suggesting a poor correspondence between the free-exploratory paradigm and the elevated plus-maze. In conclusion, the data presented here indicate that the premise of correlation between trait and state anxiety is not necessarily true for animal models of anxiety and, therefore, care must be exercised when using state anxiety models in order to determine animals' anxiety profile.