Practical considerations in establishing reliable and sensitive neurobehavioral test methods.

The use of animal models in the safety assessment of compounds requires that the methods employed be established as reliable, sensitive and valid indicators of toxicity. Neurobehavioral toxicology is in a unique position due to recent efforts in this area prior to incorporation into routine screening systems. Aspects of study design and general implications of results are discussed from the Collaborative Behavioral Teratology Study, a large-scale effort specifically undertaken to evaluate the reliability and sensitivity of a standardized behavioral teratology test protocol. In control animals in any given measure used, comparable trial-, day- or age-related patterns were obtained in all 6 participating laboratories. In addition, prenatal treatment-related alterations observed were consistent within and across laboratories in both the amphetamine and methylmercury studies. For the majority of endpoints monitored, coefficients of detection calculated from vehicle control data estimated that a 5-25% change in behavior in treated groups could be detected statistically. Thus, with use of an adequate design and sample size within a standardized protocol, it is possible to identify the presence or absence of behavioral change with some confidence.