Common And Distinct Effects Of Oxytocin On Reinforcement Learning Under Stable And Volatile Associations

As an important way for individuals to interact with the environment,reinforcement learning helps them to better adapt and survive in a changing world.Impaired ability of reinforcement learning is closely related to psychiatric disorders such as anxiety and depression.For example,anxious individuals tend to overestimate the possibility and severity of negative outcomes and they cannot timely adjust their future behavior choices based on action-outcome contingencies.One of the main problems is the dysfunction of associative learning.Therefore,it is of great application value to explore how to improve reinforcement learning ability and intranasal oxytocin（OT）may be an intervention with positive potential.As a neuropeptide,a large number of intranasal OT studies have been accumulated in the past ten years,revealing its important role in modulating individual social behavior and emotional processing.A small number of studies have explored the modulatory effect of intranasal OT on reinforcement learning in humans,but the experimental tasks of those studies are usually in a context where the action-outcome contingency is stable.Because the association between things in real life is not always consistent,it has higher ecological validity and realistic value to examine whether OT has a modulatory effect on associative learning under the volatile context and its similarities and differences with that under the stable context.Based on the above background,this study modified a classical associative learning task to explore the modulatory effect of intranasal OT on reinforcement learning in the volatile context of reward probability and its similarities and differences with that under the stable context and combined with computational modeling and event-related potential technology to reveal its more detailed behavioral patterns and neural mechanisms.Using a randomized double-blind placebo-controlled between-subject design,this study recruited a total of 80 healthy male college students to participate in the experiment.After45 minutes of intranasal OT or placebo,subjects were asked to complete a learning task involving stable and volatile association contexts.The results showed that at the behavioral level,the main effect of drugs was significant,with OT significantly improving the subjects’ choice accuracy of the optimal shape（high-probability reward shape）across the stable and volatile contexts;The main effect of context was also significant,with the choice accuracy of the optimal shape in the stable context was significantly higher than that in the volatile context.After fitting and statistical analysis of behavioral data using a reinforcement learning model of reward and punishment,it was found that OT significantly improved the subjects’ learning rates compared with placebo in the volatile context and the subjects’ choice consistency in the stable context.At the neural level,OT reduced the amplitude of P300 related to attentional resource allocation and increased N2 pc associated with selective attention in both stable and volatile contexts,but no significant effect of OT was found on ERN and FRN in either context.Correlation analysis showed that trait anxiety scores were negatively correlated with subjects’ choice consistency in both stable and volatile contexts.The above results indicate that OT effectively promote associative learning through the subjects’ choice consistency in the stable context and the learning rates in the volatile context respectively.This promotion effect is closely related to OT can improve the efficiency of attentional processing and selective attention to the optimal shape during learning.