Why stronger prefrontal synapses make the brain sleep harder
The prefrontal cortex is a good place to test this because it is heavily involved in the kind of waking activity that builds up sleep pressure. A mathematical model built for the study predicted that stronger synaptic connections push neurons toward their quiet "down state," which in turn raises delta power, the large slow waves seen in non-REM sleep [4]. In other words, a more strongly wired cortex may be easier to push into the synchronized quiet state that shows up on an EEG as slow-wave activity.
To test this causally, the researchers developed a molecular tool that enlarges dendritic spines and potentiates excitatory synapses in the prefrontal cortex on demand. When they switched it on, mice showed increased non-REM sleep amounts and higher delta power [4]. Because the manipulation targeted prefrontal excitatory neurons specifically, the result points to prefrontal synaptic strength as a causal lever on sleep pressure rather than just a correlate of it.
What slow-wave sleep does to synapses in return
The other half of the loop comes from work on synaptic downscaling. In rats, sleep reduced levels of AMPA receptors, the main receptors carrying excitatory signals at glutamatergic synapses, in both the hypothalamus and the neocortex [1]. The strongest decreases were seen for receptors containing the GluA1 subunit. By comparing selective REM sleep, total sleep deprivation, and normal sleep, the study found that slow-wave sleep was the driving force behind this down-regulation, with no additional contribution from REM sleep or the circadian rhythm [1].
One detail matters for how neatly this fits the prefrontal story: the downscaling was not linked to EEG slow-wave activity itself, even though slow-wave sleep drove it [1]. That suggests the renormalization process travels with the sleep state rather than scaling moment-to-moment with delta power. The same study also found that higher spindle density during slow-wave sleep predicted relatively higher GluA1 levels in hypothalamic synapses, which the authors connect to spindles' role in memory consolidation [1].
Inhibitory synapses and development add nuance
Excitatory synapses are not the whole story. In the cortex, microglia, the brain's resident immune cells, regulate inhibitory synapses in an activity-dependent way: sleep induced a microglia-dependent enrichment of GABA-A receptors, the main inhibitory receptors, through a pathway involving microglial TNFα and P2RX7 signaling [5]. When microglia were depleted of TNFα, slow waves during non-REM sleep changed and memory consolidation in sleep-dependent learning tasks was blunted [5]. This shows that slow waves are shaped by inhibitory as well as excitatory synaptic plasticity, and that the immune system is part of the machinery.
Development adds another layer. In adolescent mice, slow-wave activity and synaptic markers in the frontal cortex change together over the course of adolescence, and the authors note that synaptic strength rises during wake and falls during sleep in adult mice [3]. That developmental framing is consistent with the idea that the relationship between synaptic strength and slow-wave activity is not fixed but shifts as the brain matures, though the study's focus on the frontal cortex means it does not directly test the prefrontal manipulation question.
What this means for the feeling of needing sleep
Putting the pieces together, the evidence supports a two-way relationship rather than a one-way cause. Potentiating prefrontal excitatory synapses increases non-REM sleep and delta power [4], while slow-wave sleep drives downscaling of excitatory receptors in the hypothalamus and neocortex [1]. That is the shape of a homeostatic loop: waking learning strengthens synapses, the stronger synapses raise sleep pressure and slow-wave activity, and slow-wave sleep then weakens synapses again.
The loop is not complete in the evidence, though. The downscaling study did not examine the prefrontal cortex or measure sleep need directly [1], and the prefrontal manipulation study did not track receptor levels after sleep [4]. The inhibitory-synapse work shows that slow waves can be sculpted by microglial regulation of GABA-A receptors, which is a different synaptic route than excitatory potentiation [5]. So the most defensible summary is that prefrontal synaptic strength can set sleep pressure and delta power, and slow-wave sleep can renormalize synaptic strength, but the exact circuit connecting the two in the prefrontal cortex is still being mapped.
Sources used in this answer
Slow-wave sleep drives sleep-dependent renormalization of synaptic AMPA receptor levels in the hypothalamus
Slow-wave sleep drives down-regulation of AMPA receptor levels, especially GluA1-containing receptors, in the hypothalamus and neocortex of rats, independent of EEG slow-wave activity.
The Need of Slow Wave Activity and Cognitive Functions
Sleep deprivation is associated with a rebound in frontal delta activity and with downscaling of synaptic strength, discussed in relation to cognitive function.
Developmental patterns of sleep slow wave activity and synaptic density in adolescent mice
Slow-wave activity and synaptic density in the mouse frontal cortex change together over adolescence, consistent with synaptic strength rising during wake and falling during sleep.
Prefrontal synaptic regulation of homeostatic sleep pressure revealed through synaptic chemogenetics
Chemogenetic potentiation of excitatory synapses in the mouse prefrontal cortex increases non-REM sleep amounts and delta power, establishing a causal link between prefrontal synaptic strength and sleep pressure.
Microglial TNFα controls daily changes in synaptic GABAARs and sleep slow waves
Microglial TNFα regulates sleep-induced enrichment of GABA-A receptors at cortical inhibitory synapses, and depleting microglial TNFα alters NREM slow waves and blunts sleep-dependent memory consolidation.