Detection of prenatal effects on learning as a function of differential criteria

The role of appropriate criteria for detecting prenatal effects on learning was evaluated in Crl:COBS CD (SD)BR rats using a two-choice spatial discrimination escape paradigm. Alcohol was the prenatal treatment, as it has been reported to produce learning deficits in rats. The offspring of dams which consumed either a lab chow diet or isocaloric liquid diets containing 0%, 17.5% or 35% ethanol-derived calories during pregnancy were reared by surrogate control dams and tested for watermaze learning at 20 days postparturition. Learning performance and the required number of trials to the selected criterion were interrelated: (1) four or five consecutive errorless trials were too difficult, as high error rates occurred in all groups, including the control; (2) two consecutive errorless trials, or the number of trials before the first errorless trial, were too simple, as low error rates occurred in all groups, including the high dosage group; and (3) three consecutive errorless trials revealed significant dosage-dependent learning decrements in the alcohol-exposed groups. The results indicate that selection of the appropriate criterion in learning paradigms is critical for valid testing of prenatal effects.