Psychedelics do not always enhance cognitive flexibility: The potential role of time-varying effects on reward learning processes

Psychedelics may enhance cognitive flexibility, but findings are mixed, with some studies showing impairment. This review proposes that acute effects of psychedelics boost learning from reward receipt, which can lead to perseveration, while post-acute effects enhance learning from reward omission, counteracting that perseveration. Different serotonin receptor subtypes (5-HT2C, 5-HT2A, 5-HT1A) may support these distinct learning processes. This hypothesis explains when psychedelics improve flexible decision-making and why insights reached under their influence can be held strongly, and it suggests strategies to reduce maladaptive belief formation.

Study at a glance

Abstract

Although research on psychedelics (hallucinogenic serotonin 2 A or 5-HT2A agonists) has shown initial promise in ameliorating a range of psychiatric conditions, the mechanisms by which psychedelics promote therapeutic benefits have yet to be elucidated. Researchers have posited enhancements in cognitive flexibility as a transdiagnostic mechanism through which psychedelics could disrupt repetitive, maladaptive thoughts and/or behavioral patterns and induce insights. However, the acute and post-acute effects of psychedelics on cognitive flexibility have been mixed, including several cases in which cognitive flexibility is impaired. Here, we review the acute and post-acute effects of psychedelics on cognitive flexibility, specifically reversal learning and set-shifting, and propose that time-varying effects on learning from reward receipt and reward omission can account for these mixed findings. The synthesized evidence suggests that the acute effects of psychedelics may especially enhance learning from reward receipt that can become difficult to overcome, thereby leading to perseveration. In contrast, enhancements in learning from reward omission may become more prominent post-acutely and can counteract such perseveration. Whereas 5-HT2C activation appears to especially support learning from reward receipt, 5-HT2A and 5-HT1A activation may be more important for learning from reward omission. We end with proposing how different learning systems could support these two forms of learning, recommendations for future work, and implications for psychedelic therapy. Together, our hypothesis accounts for conditions in which psychedelics enhance flexible decision-making and why insights reached on psychedelics may be held so strongly, and it informs strategies to potentially minimize the formation of maladaptive beliefs.