Hypothesis

During sleep, the brain couples glymphatic clearance with activity‑dependent synaptic tagging to decide which synapses persist. We propose that slow‑wave sleep drives a surge of cerebrospinal fluid that not only flushes extracellular waste but also reaches perisynaptic spaces, allowing ubiquitin‑ligases to access synapses marked for removal by low CaMKII activity. High‑activity synapses retain protective phosphorylation tags (e.g., PSD‑95 sumoylation) that resist ubiquitin‑mediated degradation, while low‑activity synapses become ubiquitinated and are engulfed by perivascular microglial processes facilitated by CSF flow. Thus, sleep acts as a selective editing step: only synapses that met a waking‑activity threshold survive to the next day. Chronic disruption of slow‑wave sleep uncouples CSF flow from tagging, causing either excess removal of functional synapses or failure to eliminate maladaptive connections, contributing to cognitive rigidity or neuropsychiatric risk.

Mechanistic rationale

• Slow‑wave sleep lowers extracellular norepinephrine, expanding interstitial space by ~60% and doubling CSF influx 1.
• This CSF surge reaches perisynaptic clefts, delivering proteases and ubiquitin‑conjugating enzymes that can modify synaptic proteins 2.
• Neuronal activity during wakefulness sets a molecular tag: CaMKII autophosphorylation marks strong synapses; low CaMKII leaves synapses exposed to ubiquitination by activity‑regulated ligases (e.g., Ube3A) 3.
• Microglia in a surveillance state preferentially phagocytose ubiquitinated synapses when CSF flow delivers complement proteins (C1q, C3) that tag debris for clearance 4.
• Concurrently, autophagy removes damaged organelles within neurons, a process that is region‑specific and sleep‑dependent 5, preventing intracellular accumulation that could spill over into extracellular space.

Testable predictions

1. Optogenetic enhancement of slow‑wave oscillations in mice will increase CSF tracer influx into the hippocampal stratum radiatum and raise ubiquitination of PSD‑95 at synapses exhibiting low CaMKII activity, measured by proximity ligation assay.
2. Genetic knock‑down of Ube3A in excitatory neurons will abolish sleep‑dependent loss of low‑activity synapses without affecting overall CSF flow, leading to persistence of silent synapses and impaired reversal learning.
3. Pharmacological blockade of CSF influx (e.g., via acetazolamide to reduce CSF production) will prevent ubiquitin‑dependent synaptic loss despite normal slow‑wave EEG, resulting in accumulation of extracellular tau fragments as detected by ELISA.
4. In humans, high‑density EEG combined with intrathecal CSF sampling after a night of selective slow‑wave deprivation will show a negative correlation between slow‑wave power and CSF‑soluble ubiquitinated synaptic proteins (e.g., ubiquitin‑C‑terminal hydrolase L1) in young adults.

Potential falsifiers

• If boosting slow‑wave sleep fails to alter synaptic ubiquitination levels, the coupling hypothesis is weakened.
• If Ube3A deficiency does not change synapse number but rescues cognitive deficits, alternative mechanisms dominate.
• If CSF blockade does not affect synaptic tagging outcomes, then glymphatic flow may be permissive rather than instructive.

Implications

Linking synaptic tagging to glymphatic clearance reframes sleep not merely as a waste‑removal period but as an active circuit‑editing window. Dysregulation could underlie disorders where synapse number or strength is pathologically altered (e.g., autism, schizophrenia, Alzheimer's). Therapeutic strategies that enhance slow‑wave oscillations or modulate synaptic ubiquitin ligases might restore the brain’s nightly editing fidelity.