Why sleep's plasticity boost follows the networks you used, not the hours you put in
The most direct evidence comes from a motor-sequence study that trained people either with physical practice or with motor imagery (mentally rehearsing the movement) and then compared overnight sleep against staying awake [1]. After sleep, the alpha-oscillation network (a sensorimotor-premotor circuit) showed increased neural desynchronization, while the beta-oscillation network (a cortico-cerebellar circuit) showed the opposite pattern after wake consolidation. In other words, sleep did not uniformly dial plasticity up or down; it shifted activity in specific, task-relevant circuits in a direction tied to the learning context.
That pattern fits a broader picture in which sleep-related oscillations change locally in the brain areas that were used and involved in learning during wakefulness, and those local changes then modulate experience-dependent adaptation and memory consolidation [3]. The plasticity signal is therefore better described as spatially targeted than as a global quantity that scales with how much was learned.
Learning 'primes' synapses so that sleep can finish the job
A useful way to think about the relationship is priming: waking experience appears to set synapses up for structural change during subsequent sleep, so that the two brain states work in concert to produce lasting circuit changes [5]. On this view, prior learning is a precondition for sleep-related plasticity rather than a dial whose setting determines the size of the effect.
At the molecular and cellular level, sleep and sleep deprivation alter signaling pathways that regulate synaptic strength and plasticity-related gene transcription, and REM and NREM sleep can enhance previously induced synaptic potentiation, though synaptic de-potentiation during sleep has also been observed [2]. Notably, that same review describes coordinated replay of prior learning-related firing patterns during NREM sleep, and analogous learning-associated hippocampal reactivation in humans [2]. Replay is a mechanism by which sleep could act specifically on what was learned, which is consistent with a targeted rather than a proportional relationship.
The direction of the change depends on the marker and the sleep state
One reason a simple 'more learning, more plasticity' rule does not hold is that different plasticity markers move in different directions. In the motor-imagery study, the alpha network's desynchronization increased after overnight sleep, while the beta network's synchronization increased after wake consolidation instead [1]. A reader looking for a single plasticity number would see these as contradictory; read as circuit-specific signatures, they are not.
The review evidence adds that sleep can enhance previously induced potentiation but that de-potentiation during sleep has also been observed [2]. So even the sign of the effect is not fixed across markers, brain regions, or sleep stages, which makes proportionality to prior learning an unlikely description of the underlying biology.
What this means if you are trying to learn something
The practical implication is that sleep is most usefully treated as part of the learning session rather than as a separate booster whose benefit you can estimate from study volume. Because the plasticity changes concentrate in the circuits engaged during practice [1][3], what you do before sleep is likely to shape where sleep's effects land, and priming of synapses by waking experience is presented as the route through which sleep produces lasting change [5].
How much of the effect is strictly proportional to the amount learned remains an open question, and the reviewed evidence does not settle it: the findings establish that sleep engages plasticity mechanisms tied to prior learning, not a measured dose-response curve [2][5]. Separately, derailment of NREM sleep homeostatic plasticity has been linked to epilepsy, framed as a cost of the brain's plasticity [4] — a reminder that these sleep-plasticity mechanisms are powerful and not cost-free, though that work concerns pathology rather than learning dose.
Sources used in this answer
Brain plasticity underlying sleep-dependent motor consolidation after motor imagery
After motor-sequence training by physical practice or motor imagery, overnight sleep increased alpha-network desynchronization relative to wake consolidation, while the beta network showed increased synchronization after wake consolidation instead.
Sleep, plasticity and memory from molecules to whole-brain networks.
Sleep and sleep deprivation bidirectionally alter molecular pathways regulating synaptic strength, sleep can enhance previously induced synaptic potentiation (with de-potentiation also observed), and NREM sleep shows replay of prior learning-related firing patterns.
Local sleep: A new concept in brain plasticity
Local changes in sleep-related oscillations occur in brain areas used during wakeful learning and directly modulate experience-dependent adaptation and memory consolidation.
Derailment of Sleep Homeostatic Plasticity Affects the Most Plastic Brain Systems and Carries the Risk of Epilepsy
Derailment of NREM sleep homeostatic plasticity is linked to epileptogenesis in major epilepsy syndromes, framing epilepsy as a cost of brain plasticity.
Primed to sleep: the dynamics of synaptic plasticity across brain states
Waking experience is proposed to prime synapses for sleep-related structural changes, with plasticity distributed across wake and sleep states working in concert.