Central CRF and acute stress differentially modulate probabilistic reversal learning in male and female rats

• Published in: Behavioural brain research 2021-01, Vol.397, p.112929-112929, Article 112929
• Main Authors: Bryce, Courtney A., Floresco, Stan B.
• Format: Article
• Language: English
• Subjects: Cognitive flexibility; Corticotropin releasing factor; Probabilistic reinforcement learning; Sex differences; Stress
• ISSN: 0166-4328; 1872-7549
• DOI: 10.1016/j.bbr.2020.112929

•Central CRF infusion reduces negative feedback sensitivity.•CRF infusion, and to a lesser extent, acute restraint stress impairs motivation, particularly following a reversal.•Neither CRF infusion nor acute restraint stress significantly alters probabilistic reinforcement or flexibility.•Across tasks and treatments females are less motivated and less sensitive to feedback than males. Acute stress can have variable and sometimes sex-dependent effects on different executive functions, including cognitive flexibility, some of which may be mediated by increased corticotropin releasing factor (CRF). Previous studies on the effects of stress and CRF on cognitive flexibility have used procedures entailing deterministic rewards, yet how they may alter behavior when outcomes are probabilistic is unclear. The present study examined how acute stress and increased CRF activity alters probabilistic reversal learning (PRL) in male and female rats. Rats learned to discriminate between a ‘correct’ lever rewarded on 80 % of trials, and an “incorrect” lever delivering reward on 20 % of trials, with reward contingencies reversed after 8 consecutive correct choices. Separate groups received either intracerebroventricular infusions of CRF (3 μg) or restraint stress prior to a PRL session. Experiments examined how these manipulations affected learning when given prior to a one-day acquisition test or during performance in well-trained rats. Exogenous CRF, and to a lesser extent acute stress, impaired motivation across sexes, slowing deliberation times and increasing the number of trials omitted, particularly following a switch in reward contingencies. Neither manipulation significantly altered errors or reversal performance. However, increased CRF activity reduced negative feedback sensitivity. Across manipulations, females showed increased omissions and choice latencies, and were less sensitive to feedback than males. These results reveal the complexity with which stress, CRF, sex, and experience interact to alter aspects of motivation and probabilistic reinforcement learning and provide insight into how CRF activity may contribute to symptoms of stress-related disorders.