Does sleep weaken all synapses globally or only strengthened ones, and how do studies tell these apart?

Why would sleep target strengthened synapses rather than all of them?

The leading idea is that waking life builds synapses up and sleep brings them back into balance. As the brain learns and experiences the day, connections strengthen; if that growth went unchecked, the theory goes, neural circuits would saturate and become unstable. Sleep is proposed as the period when that accumulated strengthening is scaled back, serving neuronal homeostasis — keeping activity within a workable range [1].

The key word is 'strengthened.' The hypothesis is framed around synapses that gained strength during wakefulness, not every synapse in the brain [1]. That distinction matters: a global wipe would erase useful learning along with the excess, whereas scaling back the connections that grew most leaves the relative pattern of learning intact while reducing overall load.

Sleep isn't purely a weakening state

It would be a mistake to picture sleep as a simple drain on synaptic strength. Sleep also strengthens some connections. Selectively strengthening synapses during sleep has been proposed as part of what sleep does, and activity in the sleeping brain can support strengthening even while weakening occurs elsewhere [3].

Work on REM-like sleep makes the two-directional picture concrete: reactivation during that state can either weaken an excitatory connection or strengthen one, depending on the specific firing order among neurons [4]. So the same sleep state can push different synapses in opposite directions, which is hard to reconcile with a blanket global weakening.

How do researchers tell global weakening from selective weakening?

The core method is to follow individual synapses over time rather than averaging the whole brain. If sleep weakened everything globally, you would expect a broad, uniform decline. If it targets strengthened synapses, the decline should concentrate on the connections that grew during wakefulness — and the evidence is read that way, with strengthened synapses tracked as the ones subject to scaling [1].

Computational models add a second lever. By simulating different firing patterns and plasticity rules across the sleep-wake cycle, researchers can ask which rules reproduce the observed synaptic changes, and the models address both weakening during sleep and how to distinguish strengthened synapses [2]. That lets them test competing explanations rather than just describe an average shift.

A third angle is to look at where on the synapse change happens. Sleep appears to modulate strength differently in the presynaptic and postsynaptic compartments, with weakening during sleep reported alongside changes in spine density [5]. Compartment-level measurements can therefore reveal selective effects that a single whole-synapse number would blur.

What is still being worked out

The picture is suggestive rather than settled. Much of the support comes from models and from hypotheses about scaling, and the proposal that sleep exists to rebalance strengthened synapses remains a hypothesis rather than a demonstrated purpose [1][2]. The two-directional findings also mean the net effect on any given circuit depends on which synapses are engaged, so 'sleep weakens synapses' is too blunt to be accurate [3][4].

What the evidence does support is a shift in the question: not whether sleep weakens synapses globally, but which synapses it weakens, which it strengthens, and under what firing conditions. The tools for answering that — tracking individual connections, modeling plasticity rules, and measuring pre- and postsynaptic compartments separately — are the same tools that revealed the selectivity in the first place [2][5].

Sources used in this answer

Related research questions