Why does training near bedtime seem to stick better?
The clearest evidence comes from a study that trained three groups on a finger-tapping task at 10 a.m., 3 p.m., or 8 p.m. and retested them 24 hours later. Immediate skill gain right after training was the same regardless of the hour, so the clock did not change how well people learned in the moment. What differed was what happened afterward: the evening group improved over the next day, the afternoon group held steady, and the morning group got worse [1].
The authors ran two control experiments to test whether sleep itself was doing the work. One group trained at 8 p.m. but stayed awake, and another trained at 8 p.m. and slept; the comparison supported the idea that a night of sleep contributes to the consolidation seen in the evening group [1]. In plain terms, the skill practiced shortly before bed appears to get processed during sleep, while a skill practiced in the morning sits through a long waking day before the retest and tends to decay.
If sleep is the point, why does afternoon training also consolidate well?
A second study complicates a simple "sleep is everything" story. It trained people in the morning (10 a.m.) or afternoon (3 p.m.), using either physical practice or mental rehearsal (imagining the movements), and retested 24 hours later. Everyone improved right after practice, and physical practice produced bigger immediate gains than mental rehearsal. But when it came to holding onto the skill overnight, both the physical and the mental groups did better when training happened in the afternoon [2].
That result points to a genuine time-of-day effect on consolidation that is separate from sleep timing. The authors conclude that time of day positively influences motor skill consolidation and should be considered when scheduling training in sport or clinical settings [2]. Read alongside the evening-training result, the pattern suggests consolidation is best when practice lands in the later part of the day, not that any single hour is magic.
Does sleep help every kind of motor practice the same way?
No. A study of whole-body sequence learning compared physical practice with motor imagery practice, and compared a night of sleep with an equivalent daytime period. Both practice groups improved during training. But only the imagery group that slept showed further improvement after the night; the physical-practice group and the daytime groups stabilized rather than gained [3]. The authors describe this as sleep contributing to consolidation of gross motor sequence learning acquired through motor imagery, but not through physical practice [3].
That is a striking qualification. It means "overnight gains" is not a single phenomenon: the type of practice can change whether sleep adds anything at all. The same study found that the goal-related and movement-related aspects of the task changed together, without being affected by practice type or by whether consolidation happened over sleep or daytime [3], so the sleep benefit showed up in overall performance rather than in one component of the movement.
So what should you conclude when a retest shows improvement?
An overnight improvement is consistent with sleep-dependent consolidation, and at least one line of work reports that improvements occur across a night's sleep on a sequential finger-tapping task [4]. But the time-of-day studies show that when you trained matters too, and that morning practice can lead to forgetting rather than gain even when sleep follows [1]. A retest gain therefore reflects a mix of sleep-related processing and the circadian timing of both the training session and the retest, not one factor alone.
The practical upshot, as the authors of these studies frame it, is that scheduling matters: training placed later in the day, closer to sleep, appears to favor retention, and this has been proposed as relevant for planning sports, clinical, and experimental training programs [1][2]. The evidence is still mostly from small laboratory tasks like finger-tapping and whole-body sequences, so how strongly this carries over to real-world skill learning is not settled by these studies.
Sources used in this answer
Time of day and sleep effects on motor acquisition and consolidation
Training on a finger-tapping task at 8 p.m. led to performance improvement 24 hours later, while 10 a.m. training led to deterioration and 3 p.m. training stabilized, with control experiments supporting a contribution of nighttime sleep.
Time-of-day effects on skill acquisition and consolidation after physical and mental practices
Afternoon (3 p.m.) training produced better 24-hour consolidation than morning (10 a.m.) training for both physical and mental practice on a finger-tapping task, indicating a time-of-day effect on consolidation.
Sleep dependent consolidation of gross motor sequence learning with motor imagery
In gross motor sequence learning, only the motor-imagery group that slept improved further after a night of sleep; physical-practice and daytime groups stabilized.
Sleep and the time course of motor skill learning
Sequential finger-tapping performance improved across a night of sleep, supporting a role for sleep in motor skill consolidation.
Time-of-day effects on motor learning
Time of day affects motor skill acquisition and consolidation, with a night of sleep contributing to consolidation and time of day influencing the consolidation process.