Computational Analysis of Psilocybin Effects on Three-Choice Touchscreen Reversal Learning in Rats: A Pilot Study

Introduction: Cognitive flexibility is essential for behavioral adaptation in response to environmental changes and is impaired in various neuropsychiatric disorders. The serotonergic psychedelic psilocybin has shown potential in enhancing cognitive flexibility, although with mixed results. In this study, we investigated the effects of psilocybin on cognitive flexibility in rats using a three-choice visual reversal learning task.

Methods: Sixteen Long-Evans rats were trained on a three-choice touchscreen-based reversal learning task. Following training and initial reversal testing, psilocybin (1 mg/kg, subcutaneous) or saline was administered in a crossover design. Behavioral performance was assessed through recording the time to achieve learning criteria, reaction times, and interaction patterns with image choices, focusing on short-term effects (days after administration during the crossover interventions) and trends over subsequent reversals (weeks). Using computational reinforcement learning models, we analyzed development in latent cognitive parameters, such as learning rate and value sensitivity.

Results: Five rats successfully completed all six reversal learning protocols, demonstrating significant learning and unlearning dynamics. The number of sessions to attain learning criteria and the reaction time for the touchscreen decreased significantly across protocols. Psilocybin administration impaired the rate of learning and unlearning across sessions during the first postdrug reversal. Computational modeling identified low learning rates, with no significant differences between psilocybin and placebo conditions on any parameters. However, exploratory analysis revealed enhanced rates of learning and unlearning over the session in the second postdrug reversal compared with baseline.

Conclusions: Only a subset (31%) of rats effectively engaged in this complex three-choice visual reversal learning task, thereby demonstrating learning and unlearning over time. Psilocybin impaired short-term performance in learning and unlearning speeds, whereas exploratory findings suggested possible long-term enhancements in learning dynamics. Results show a nuanced effect of psilocybin on cognitive flexibility, with potential relevance for its interventional use in neuropsychiatric disorders. Further research should explore long-term outcomes and refine computational models for complex tasks.