Unraveling the Role of REM and Non-REM Sleep in Visual Learning and Brain Plasticity

Post by Elisa Guma

What's the science?

Sleep is important for learning and memory, amongst many other vital functions. However, there are still some open questions regarding the underlying neural mechanisms of sleep: (1) how do different phases of sleep such as non-rapid eye movement (NREM) or REM sleep, influence learning and memory? (2) is learning acquired before sleep enhanced or stabilized after sleep, or does sleep protect what was learned from being “overwritten” by learning something new? (3) is the facilitation of learning learning-specific or learning-independent? This week in Nature Neuroscience, Tamaki and colleagues addressed these questions by measuring the degree of plasticity and stability in neural signals during different phases of sleep after a visual perceptual learning task.  

How did they do it?

The authors trained participants on two different visual texture discrimination tasks (designed to interfere with each other), one before and one after sleep, and then tested performance on both after sleep. During sleep, they recorded glutamate and GABA (the main excitatory and inhibitory neurotransmitters in the brain) in visual areas using magnetic resonance spectroscopy (MRS) (a non-invasive technique for measuring brain metabolites). They simultaneously acquired polysomnography, a method used to record brain waves to identify sleep stages and cycles. Simultaneous MRS and polysomnography were also repeated in a separate session with no learning in order to investigate the E/I balance during different sleep phases without learning.

The ratio between GABA and glutamate was used to calculate an excitatory/inhibitory (E/I) balance, which is thought to be a reliable index of plasticity and stability in the brain. Thus, a high E/I balance indicates increased plasticity, and a low E/I balance indicates increased stability. This allowed authors to compare E/I balance (i.e. plasticity) in different phases of sleep (NREM and REM relative to wakefulness), and in individuals who had both NREM and REM or just NREM sleep. Further, the authors wanted to know if enhancement in plasticity for pre-sleep learning is followed by a less plastic period of stabilization and whether learning of one task interfered with the learning of another. They also examined whether these changes were specific to learning or simply innate to sleep.

What did they find?

Plasticity — based on E/I balance — of visual areas was increased during NREM sleep, and predicted improvements in performance on the learned task. Conversely, the E/I balance decreased in REM. The authors also examined whether pre-sleep learning stabilization occurred during sleep and found that individuals who had both NREM and REM sleep were able to maintain their learning from the pre-sleep training without interference from the post-sleep training on the new task relative to those who only had NREM sleep. The degree of resilience was also negatively correlated to E/I balance. 

The authors then compared E/I balance in the different sleep phases when pre-sleep learning had occurred, to when no pre- or post-sleep training was performed. They found that E/I balance during NREM sleep was increased in both scenarios, irrespective of pre-sleep learning. Conversely, REM sleep E/I balance was only decreased from baseline after pre-sleep learning; if no learning had occurred, it was no different than baseline. This suggests that REM sleep played a stabilizing role, protecting the learning before sleep from interference by new learning after sleep, whereas NREM sleep increases plasticity in a learning-independent manner. After taking a closer look at the neurotransmitters, GABA decreases seemed to be driving changes in E/I balance during NREM sleep, whereas glutamate decreases seemed to drive E/I balance decreases during REM sleep after learning.

What's the impact?

This study found there are two distinct mechanisms for NREM and REM sleep, with complementary patterns of neurochemical and functional processing, involved in learning facilitation. During NREM sleep, plasticity increases independent of whether or not learning occurred, whereas stabilization occurs during REM sleep only if pre-sleep learning occurred. While these findings aid in our understanding of the relationship between sleep and learning, future work may be extended to understanding how sleep abnormalities may interfere with learning and memory.

Tamaki et al. Complementary contributions of non-REM and REM sleep to visual learning. Nature Neuroscience (2020). Access the original scientific publication here