Methodological problems in the study of classical aversive conditioning of adrenocortical responses

The present studies were conducted to demonstrate classical aversive conditioning of a corticosterone stress response in male rats. In the first experiment animals exposed to an odor which had previously preceded tube restraint stress had significantly higher plasma corticosterone levels than animals which were probed with a novel odor, or animals in which stress was followed, rather than preceded, by the odor. Careful consideration of corticosterone responses obtained on different days during training made it impossible to conclude unequivocally that learning had occurred. A second experiment was designed specifically to deal with the problems raised by the first, but we were again unable to show rigorously that learning had taken place. However, with data and paradigms comparable to our own, other researchers have prematurely concluded that aversive visceral conditioning can occur. It is suggested that the methodological considerations raised in the present studies could also be important for interpreting other experiments. Given the classical importance ascribed to learning components in many psychosomatic pathologies, future experiments are necessary to determine if stress responses can be classically conditioned, but they must be more rigorously designed than in the past.