Is sleep the price we pay for learning?

One proposed answer, from Giulio Tononi and Chiara Cirelli, is yes: the synaptic homeostasis hypothesis (SHY) proposes that sleep is the price the brain pays for plasticity. According to SHY, learning while awake means strengthening connections throughout the brain. That raises cellular needs for energy and supplies, lowers signal-to-noise ratios and saturates learning. During sleep, they propose, spontaneous activity renormalizes net synaptic strength and restores cellular balance.

The proposal has a particular view of memory. SHY says synapses of circuits that were repeatedly strengthened while awake, or that fit with older memories, are protected, while those of rarely used circuits are progressively weakened and eventually eliminated over many wake/sleep cycles.

There is evidence. In rats, levels of GluA1-containing receptors, whose movement in and out of synapses is considered a primary mechanism of strengthening and weakening them, are 30 to 40% higher after wakefulness than after sleep. In humans, the response of frontal cortex to magnetic stimulation increases over 18 hours of wake and returns to baseline after a night of recovery sleep. In fruit flies, the number of dendritic spines on a visual neuron grows after 12 hours of wake in an enriched environment and shrinks again only if the flies are allowed to sleep.

But it is a hypothesis with rivals. The active system consolidation hypothesis says that repeated reactivations of new information in the hippocampus during non-REM slow oscillations mediate stabilization and integration of memory into cortical networks. Both models agree that activity during sleep drives plasticity. What remains controversial is the direction of synaptic changes during sleep, potentiation or depression.