Testing for behavioral effects of agents.

Abstract

In the present state of science no morphological or chemical changes may be detectable at a time when behavior is profoundly disturbed, as in schizophrenia. Until we are reassured to the contrary, we must assume that exogenetic intoxication can produce changes detectable only as behavioral changes. Therefore behavioral toxicology must be studied. In contrast to toxic manifestations such as lethality or carcinogenicity, which tend to be unequivocal and irreversible, behavioral changes are like physiological changes in that they are quantitative, changing in time, and relate to variables with a considerable range of normal variability. An experiment on behavioral teratology in mice is described and the results used to illustrate the limits of the possible in behavioral toxicology. From reported and observed variability it is surmised that changes that occur in as many as 1 per 100 of the population or average as large as a 10% decrement will still be too small to be detected by direct experiment. Such risks are frequently unacceptable. Reasons are given for hoping that epidemiological studies may be able to supplement experimental toxicological studies to provide a better assessment of risk of small impairments or rare susceptibility.