Plasma cortisol response cannot be classically conditioned in a taste-endocrine paradigm in humans

Peripheral immune responses can be modified by associative learning procedures. Less is known, however, whether and to what extent neuroendocrine parameters can be classically conditioned. In this randomized double-blind study, we modified an established paradigm to behaviorally condition endocrine responses in humans. Thirty-one healthy male participants received a distinctively flavored green drink as the conditioned stimulus (CS) and intravenous injections of corticotropin-releasing hormone (CRH) (CRH group, N = 17) or NaCl (placebo group, N = 14) as the unconditioned stimulus (US) during two subsequent acquisition trials. Plasma levels of cortisol and noradrenaline, heart rate, and psychological parameters were analyzed before and 15, 30, 60, 120, and 180 min after injection. The two acquisition trials were followed by two evocation trials, during which participants underwent the same procedure but now receiving NaCl injections. CRH administration induced pronounced increases in cortisol and noradrenaline plasma concentrations, heart rate, and anxiety levels. However, re-exposure to the CS during evocations trials did not provoke conditioned increases in neuroendocrine parameters. Median split of the CRH group based on the cortisol baseline level into cort-high and cort-low subgroups showed that the cort-high subgroup displayed a significantly increased cortisol production on evocation days compared to the cort-low subgroup and the placebo group. This taste-endocrine paradigm employing CRH injection as the US in healthy male volunteers failed to induce a behaviorally conditioned cortisol release as a learned endocrine response. Future studies should clarify a possible role of higher baseline cortisol levels in perhaps facilitating a conditioned cortisol response.