Using signal detection methods for analysis of operant performance in mice.

Several factors account for murine cognitive abilities, and manipulation of genes which would act at the effector molecules involved in stimulus processing, reward-related properties and/or motor output can easily confound behavioural data obtained from mouse mutants responding on cognitive tasks. Therefore, tests may be needed which allow a better dissociation between true cognitive processes (accuracy) and other factors that may alter performance (motor or motivational bias). Part of this can be achieved by using methods which enable parametric variation of task difficulty. Part of it can also be achieved by using data analysis that allows a distinction between accuracy and bias, such as the mathematical methods of signal detection theory (SDT). SDT formally addresses the possibility that a given gene product or lack thereof affects performance by affecting motivation rather than cognition. It proposes that performance in a task depends on two factors, that is the sensitivity (or accuracy) of the neural systems mediating a cognitive process and the subject's motivational state, the latter of which can be represented as bias. SDT analysis can be easily applied to murine data. This overview will discuss the advances and limitations of the various SDT measures and illustrate the value of this type of analysis for understanding cognitive performance of mice.