Reliability, sensitivity and validity of behavioral indices of neurotoxicity

Adequate sensitivity, reliability, and validity are features of scientific testing in all fields. Accordingly, these have been central concerns in the development of behavioral methods for assessing neurotoxicity. Advances during the last 10 years have established behavioral tests that show high degrees of sensitivity and reliability both within and across laboratories. This was demonstrated in the recently completed National Center for Toxicological Research (NCTR) Collaborative Behavioral Teratology Study and other collaborative studies. Validity of behavioral testing has been based on: (a) the use of known neurotoxins; (b) comparison of agents causing behavioral dysfunction in humans; and (c) theoretical considerations of domain of central nervous system (CNS) function to be evaluated. These approaches have worked well, however, the neurochemical/neuropathologic to behavior correlations have been effective only under circumstances where a neurotoxic compound, as recognized by an identifiable neurobiologic signature, has been available. Behavioral methods are important in neurotoxicity evaluations because they provide an assessment of overall CNS functional integrity, whereas other methods in the neurosciences are directed towards specific subsystems. This presentation discusses the validity of measurements in: (1) behavioral testing; (2) neuropathologic evaluations; (3) neurochemical tests; and (4) electrophysiological techniques for detecting neurotoxicity.