The REM effect: How sleep stages influence learning strategies in complex reward-based decision-making

Area of Science:

• Neuroscience
• Cognitive Psychology
• Sleep Science

Background:

• Model-free (MF) and model-based (MB) learning are key decision-making processes.
• Compulsive behaviors may stem from overreliance on MF learning.
• The role of sleep in modulating MF/MB learning strategies is unknown.

Purpose of the Study:

• To investigate the influence of sleep manipulation on MF and MB learning strategies in healthy adults.
• To explore the relationship between sleep quality and the engagement of distinct learning strategies.

Main Methods:

• 67 healthy adults underwent a week of either sleep restriction (5 hours/night) or well-rested sleep (9 hours/night).
• Participants completed a two-stage reinforcement learning task.
• Mixed-effect logistic regressions and computational modeling were used to analyze learning strategies.

Main Results:

• No significant differences in MF/MB learning were found between sleep conditions.
• Reduced REM sleep correlated with increased MB learning.
• Increased REM sleep correlated with increased MF learning.
• Computational modeling confirmed a negative association between MB learning parameters and REM sleep percentage (τ = -0.22, p = 0.02).

Conclusions:

• REM sleep quantity may influence the dominant learning strategy employed.
• Individuals with less REM sleep might be less adept at evaluating costs and benefits of choices.
• Further research is warranted to elucidate the role of REM sleep in decision-making strategies.