Sleep's effects on beta oscillations predict the adaptation of motor behavior

The remarkable plasticity of the sensorimotor cortex enables the brain to update its knowledge and adjust movements to meet the demands of the surrounding environment, a process known as motor adaptation. In this study, we aimed at investigating the role of sleep in the adaptation of an automatic fine-motor skill. We recruited experts on touch-typing on the regular keyboard (N=33, all males) and trained them to type on a mirrored keyboard, then tested their performance after a retention interval of either overnight sleep with polysomnography (N=16), or daytime wakefulness (N=17). During training and testing, participants had to type 5-letter; German words on regular and mirrored keyboards, respectively, while we measured their brain activity via electroencephalography (EEG). We show that sleep benefits adaptive performance. Typing performance, i.e. a measure that accounts for both typing accuracy and speed, on the mirrored keyboard improved significantly after sleep but not after wakefulness. Adaptive performance-gain after sleep correlated positively with the change in the slope of the aperiodic brain activity (3-55Hz) measured over all leads. The improved adaptive performance after sleep was accompanied by a significant decrease in the power of beta oscillations (13-30Hz). However, the decrease in beta band activity after sleep occurred for both the regular and the mirrored keyboards suggesting a general increase in adaptive drive. We then trained a classifier to differentiate between regular and mirrored typing and we show that the classifier’s performance dropped significantly after sleep but not wakefulness, corroborating the role of sleep in consolidating the adaptation task and suggesting a role for sleep in inducing interference between the two tasks. Therefore, we checked whether the typing performance on the regular keyboard declined for the words that were trained on the mirrored keyboard (List 1) as compared to words that were typed only on the regular keyboard (List 3) before sleep. Indeed, we found that only after a period of sleep but not wakefulness, the typing performance on List 1 decreased significantly, while that on List 3 showed a significant increase. Our results suggest that sleep promotes the consolidation of all forms of motor behavior indiscriminately, possibly during different phases of sleep. We speculate that bursts of beta oscillations during tonic episodes of rapid eye movement (REM) sleep might reflect the consolidation of the motor aspects of adaptive behavior.