Sleep has long been described as something that happens to the cortex — the brain's outer layer, responsible for perception, reasoning, and memory, was thought to be switched off by deeper structures like the brainstem and hypothalamus once the body decides it's time to rest. A study published September 9 in Nature complicates that picture, identifying a strikingly rare population of neurons within the cortex itself that appears to actively drive the transition into sleep.
An Overlooked Cell Type, Conserved for Hundreds of Millions of Years
The research, led by Renata Batista-Brito, PhD, now an associate professor of neuroscience at the Icahn School of Medicine at Mount Sinai and previously based at Albert Einstein College of Medicine, focused on a neuron subtype called Sst-Chodl cells — named for their co-expression of the genes somatostatin (Sst) and chondrolectin (Chodl). The paper, "Neocortical long-range inhibition promotes cortical synchrony and sleep," reports that these cells make up only about 0.2% of all neurons in the cortex, making them among the rarest known cortical cell types.
What sets Sst-Chodl neurons apart structurally is their reach. Most inhibitory neurons in the cortex are local operators, communicating primarily with nearby cells. Sst-Chodl neurons instead send projections across long distances, giving this vanishingly small population an outsized ability to influence electrical activity across wide areas of the cortex simultaneously. The researchers also found that Sst-Chodl cells have been preserved across an enormous evolutionary span, appearing in amphibians and reptiles as well as mammals — a pattern that typically signals a cell type performing some fundamentally important function.
Selectively Active When the Brain Powers Down
Using recordings in mice, Batista-Brito's team found that Sst-Chodl neurons are selectively active during low-arousal states and largely silent during periods of high arousal and active wakefulness — the opposite activity pattern from many other cortical neuron types. That alone suggested a role tied specifically to rest rather than to sensory processing or movement.
The more direct evidence came from an intervention: when the researchers artificially activated Sst-Chodl neurons in awake mice, electrical activity across the neocortex slowed and became markedly more synchronized, taking on the large, coordinated wave patterns that normally characterize sleep. Critically, this activation didn't just mimic sleep-like brain activity — it was sufficient to promote actual sleep, causing mice to fall asleep more readily and to spend more total time asleep.
Rethinking Where Sleep Comes From
The finding pushes back against a model in which sleep is generated exclusively by specialized subcortical circuits and then imposed on a passive cortex. Instead, the study suggests the cortex is an active participant in its own shutdown, with a tiny, long-range-projecting neuron population capable of tipping large swaths of cortical tissue into a synchronized, sleep-like state. Sleep, in this framing, emerges from an interaction between deep brain sleep centers and the cortex itself, rather than being a one-way command from the former to the latter.
Because Sst-Chodl neurons are so evolutionarily conserved, the researchers argue that understanding their specific role could open a new angle on sleep regulation more broadly — one focused on the cortex's own machinery rather than solely on the brainstem and hypothalamic circuits that have dominated sleep neuroscience for decades.
What This Means for Patients
This is a mouse study of a single, extraordinarily rare cell type, and any clinical application is likely years away. But the discovery gives sleep medicine a new and highly specific target to investigate: a defined cortical circuit that can be manipulated to induce sleep-like brain states, rather than a diffuse process spread across the whole brain. Conditions marked by an inability to enter or sustain sleep — including chronic insomnia and some sleep disturbances associated with neurological disease — may eventually be traceable, at least in part, to dysfunction in this specific neuron population or its long-range connections, offering researchers a more precise starting point than has previously been available for developing targeted interventions.
