Reinforcement and Punishment Shape the Learning Dynamics in fMRI Neurofeedback

Introduction: Neurofeedback (NF) using real-time functional magnetic resonance imaging (fMRI) has proven to be a valuable neuroscientific tool for probing cognition and promising therapeutic approach for several psychiatric disorders. Even though learning constitutes an elementary aspect of NF, the question whether certain training schemes might positively influence its dynamics has largely been neglected. Methods: To address this issue, participants were trained to exert control on their subgenual anterior cingulate cortex (sgACC) blood-oxygenation-level-dependent signal, receiving either exclusively positive reinforcement (PR, “positive feedback”) or also positive punishment (PP, “negative feedback”). The temporal dynamics of the learning process were investigated by individually modelling the feedback periods and trends, offering the possibility to assess activation changes within and across blocks, runs and sessions. Results: The results show faster initial learning of the PR+PP group by significantly higher activations of the sgACC in the first session and stronger regulation trends during the first runs. Independent of the group, significant control over the sgACC could further be shown with but not without feedback. Conclusion: The beneficial effect of PP is supported by previous findings of multiple research domains suggesting that error avoidance represents an important motivational factor of learning, which complements the reward spectrum. This hypothesis warrants further investigation with respect to NF, as it could offer a way to generally facilitate the process of gaining volitional control over brain activity.