What replay actually looks like during a rest break
Replay is the brain reactivating a pattern of activity it just used, as if running through an experience again off-line. In the human motor study, participants practiced a sequential finger-tapping skill while their brain activity was recorded with magnetoencephalography (MEG), a method that tracks fast electrical/magnetic signals at the scalp. During the short rests between practice blocks, the trained movement sequence reappeared in neural activity in a compressed form, running about 20 times faster than the actual movements [1].
This replay was not random noise. It was selective for the sequence people had actually trained, it appeared in both forward and reverse order, and it extended beyond the hippocampus and entorhinal cortex (memory-related regions) into sensorimotor cortex on the opposite side of the body being used [1]. That spread matters: it suggests the reactivation is not confined to a memory hub but reaches the motor machinery that would execute the skill.
Why a short waking rest may be doing consolidation work
The practical implication is that rest is not dead time. In the human study, replay occurred during the very rest periods in which rapid consolidation happened, and the strength of replay predicted how much the skill improved [1]. This is consistent with the idea that the brain uses idle waking moments to bind separate action representations into a smoother, more temporally precise sequence, though the study documents a correlation with consolidation rather than proving replay causes it.
Animal work adds context for why waking replay could be well suited to this. In rats, hippocampal replay occurs during both waking and sleep, and awake replay has been observed reactivating remote experiences [2]. Recordings of hippocampal-prefrontal reactivation during learning also show that reactivation patterns differ between awake and sleep states [3]. These findings are in rats and do not address human skill, so they support the general plausibility of waking replay without settling the human question.
Does this mean sleep replay is unimportant?
No. The human skill study did not compare waking replay with sleep replay, so it cannot show that sleep replay is absent or irrelevant [1]. What it does show is that robust, fast hippocampo-neocortical replay is present during waking rest and tracks consolidation of a motor sequence. The comparison between awake and sleep replay in animals indicates both states involve reactivation, with differences in pattern and sensory context [2][3], so the reasonable reading is that waking replay is a demonstrated contributor for this kind of skill, not the exclusive mechanism.
Sources used in this answer
Consolidation of human skill linked to waking hippocampo-neocortical replay
During waking rest after practicing a sequential motor skill, humans show fast hippocampo-neocortical replay compressed about 20-fold, selective for the trained sequence and predictive of skill consolidation; the study did not compare this with sleep replay.
Awake replay of remote experiences in the hippocampus
In rats, hippocampal replay reactivates stored representations during both waking and sleep, including awake replay of remote experiences.
Hippocampal-prefrontal reactivation during learning is stronger in awake compared with sleep states
In rats during learning, hippocampal-prefrontal reactivation patterns differ between awake and sleep states.