Brainstem Neurons in Median Raphe Generate Sleep Drive in Mice

Stimulating a diverse population of brainstem neurons in the median raphe induces deep, recovery-like sleep in mice, whereas chronically silencing them slashes non-rapid eye movement sleep by approximately 70 percent, according to a brain-wide investigation conducted by Joo and colleagues. Published in Nature and highlighted in Cell Research by Charlotte E. Luff, Sarah L. Ly, and Luis de Lecea, the findings demonstrate how specific subcortical networks generate homeostatic sleep pressure. Across the animal kingdom, sleep is defined by behavioral quiescence, elevated arousal thresholds, reversibility, and homeostasis, the vital restorative response to deprivation that separates true sleep from circadian rhythms. While seminal invertebrate research demonstrated that reactive oxygen species accumulate within the mitochondria of dorsal fan-shaped body sleep-active neurons in fruit flies to alter neuronal excitability after sleep deprivation, the physical mechanisms encoding mammalian sleep drive have remained elusive.

To establish whether sleep pressure is sensed universally by all mammalian cells or governed by specialized neural populations, Joo and colleagues performed an unbiased, brain-wide activation mapping study tracking the immediate-early gene cFos across 162 mice tested under 26 distinct conditions. The researchers examined normal circadian sleep-wake cycles alongside multiple sleep deprivation paradigms and recovery periods. Their systematic screen uncovered a specific class of wake-correlated neuronal responses that scaled proportionally with the duration of wakefulness across different deprivation methods. Only two brain structures exhibited this sustained pattern of activation that increased during wakefulness and decayed during sleep: the anterior medial preoptic area in the hypothalamus and the median raphe within the brainstem.

Using targeted activity-dependent genetic tagging with TRAP2, the investigators showed that chemogenetically reactivating deprivation-responsive neurons in either the anterior medial preoptic area or the median raphe was sufficient to trigger deep, recovery-like sleep. Conversely, chronically suppressing these populations through chemogenetic inhibition or through the viral overexpression of the inward-rectifying potassium channel Kir2.1 prevented the expected accumulation of sleep pressure. When subjected to sleep deprivation protocols, mice with inhibited median raphe neurons exhibited a marked suppression of sleep drive, as measured directly by a dramatic reduction in behavioral sleep attempts.

Detailed cellular characterization within the median raphe revealed that distinct neurochemical subtypes exert opposing influences on sleep and wakefulness. GABAergic neurons, which constitute roughly 60 percent of the median raphe neuronal population, act in concert with local serotonergic neurons to actively promote sleep, whereas an adjacent glutamatergic population drives behavioral wakefulness. Following periods of sleep deprivation, the median raphe GABAergic neurons undergo an intrinsic increase in cellular excitability. This heightened excitability provides a cellular explanation for how wake-correlated brain structures alter their circuit-level activity across shifting phases of the sleep-wake cycle.

When the researchers chronically lowered the excitability of these median raphe GABAergic and serotonergic neurons using Kir2.1 channel overexpression, the animals experienced an approximate 70 percent reduction in non-rapid eye movement sleep alongside a sustained reduction in overall sleep need. Surprisingly, despite this extensive sleep reduction, the mice did not display the severe behavioral or cognitive deficits that typically follow severe sleep loss. Furthermore, lethality reached 17 percent over a 30-day monitoring period, a mortality rate that was considerably lower than anticipated for an intervention that eliminates the vast majority of non-rapid eye movement sleep.

Serotonergic neurons have historically been linked to sleep regulation, yet their precise role in homeostatic sleep pressure remained controversial because the serotonin system features extensive cellular heterogeneity, diverse receptor subtypes, and complex downstream signaling pathways. The work by Joo and colleagues distinguishes itself from earlier characterizations by demonstrating three key phenomena: reactivating deprivation-sensitive median raphe neurons directly intensifies sleep pressure, these neurons alter their intrinsic excitability in response to prolonged wakefulness, and inhibiting them dramatically diminishes physiological sleep need. These results suggest that sleep-promoting circuits are organized in a defined functional hierarchy, operating much like wake-promoting systems where lateral hypothalamic hypocretin and orexin neurons act as master switches to coordinate systemic arousal.

Despite these significant insights, key technical and conceptual questions remain regarding the exact role of median raphe signaling. Because Fos protein expression is slow and cumulative, TRAP2 genetic tagging cannot resolve the precise moment a neuron first responds to sleep loss, creating an inherent experimental bias toward capturing persistently active cells. Sustained neuronal firing during extended wakefulness could reflect either the progressive accumulation of homeostatic sleep pressure or the active neural maintenance of wakefulness itself. Although the authors showed that median raphe excitability shifts with wake duration, that stimulation induces sleep, and that inhibition alters electroencephalographic delta-power accumulation, their study did not identify the upstream molecular signals that trigger median raphe activation or explain why GABAergic neurons become more excitable.

Because silencing median raphe neurons decreased both sleep duration and sleep demand without immediately triggering classic sleep-deprivation impairments, the authors suggest these manipulated mice do not represent sleep-deprived animals, but rather individuals that accumulate less homeostatic sleep pressure. However, sleep pressure in this work was measured using electroencephalographic delta power and sleep-rebound metrics rather than direct molecular biomarkers. Longer-term evaluations and high-resolution synaptic analyses are necessary to detect potential hidden physiological costs, especially given the observed 17 percent mortality rate. Identifying the direct upstream triggers and molecular effectors of the median raphe and anterior medial preoptic area represents a crucial step toward understanding whether future therapeutics can safely mitigate excessive sleep pressure or pharmacologically replicate the restorative benefits of recovery sleep.

Subject of Research: Circuit mechanisms of mammalian sleep drive and median raphe homeostatic control

Article Title: Deciphering sleep drive in the brain: new circuit-level insights

Article References: Luff, C. E., Ly, S. L., & de Lecea, L. (2026). Deciphering sleep drive in the brain: new circuit-level insights. Cell Research. https://doi.org/10.1038/s41422-026-01317-6

Image Credits: AI Generated

DOI: 10.1038/s41422-026-01317-6

Keywords: sleep drive, neuroscience, median raphe, homeostasis, brainstem circuits, sleep deprivation, preoptic area, optogenetics, Deciphering, sleep, drive, brain