A rare long-range cortical neuron can push the mouse brain toward sleep

A mouse study identifies a sparse inhibitory neuron type that links distant cortical regions. Its activity rises during low arousal and slow-wave sleep; experimental activation increases cortical synchrony and sleep. The result establishes sufficiency in mice, not necessity, a human therapy, or a replacement for subcortical sleep circuits.

AI authorship and translation disclosure

Conceptual mechanism assembled from the cited experiments. The cover is an AI-generated conceptual illustration, not microscopy or experimental data.

1. What changed

Sleep is usually explained from the bottom up: circuits in the hypothalamus, brainstem and basal forebrain change the state of the cortex. Ratliff and colleagues add a complementary route. In mice, a genetically defined subset of somatostatin neurons expressing chondrolectin—Sst-Chodl cells—can inhibit targets over millimetres of neocortex and help synchronize slow activity.

The advance is the combination of anatomy, recording and perturbation. The authors did not merely find a sleep-correlated signal. They reconstructed axons, measured activity across behavioural states, tested synapses in slices, and activated the cells in living animals. Together these experiments support a causal statement: activating this population is sufficient to promote cortical synchronization and more sleep in the tested mice.

2. A rare cell with unusually long axons

Most cortical inhibitory interneurons act locally. Sst-Chodl cells are exceptional: they are fewer than 1% of cortical GABAergic neurons and send ipsilateral axons across several cortical areas. Sixteen reconstructed Sst-Chodl cells from three mice showed millimetre-scale projections, unlike the local comparison cells. “Long-range” here is long within a mouse hemisphere, not a projection across the whole brain.

3. Observation and causal tests

Two-photon calcium imaging linked the labelled cells to naturally occurring state changes. In superficial layers, 95 of 111 labelled cells from 15 mice were more active during quiet wakefulness or slow-wave sleep; 16 showed the opposite pattern, probably reflecting non-target or minority cells. In deep layers, 14 of 16 recorded cells from five mice followed the low-arousal pattern. Their activity rose before cortical DOWN states and tracked 1–4 Hz delta activity, but regression did not show that it predicted later delta changes.

Optogenetic activation increased low-frequency cortical power and synchrony over recording separations of roughly 2–6 mm, with a weaker effect at greater distance. In slices, recordings from 156 cells in 16 mice found responsive targets even around 2 mm from the injection site. Gabazine blocked the responses, and tetrodotoxin plus 4-aminopyridine supported a monosynaptic GABAergic connection. Polysynaptic propagation through ordinary corticocortical pathways cannot be excluded in the intact brain.

Chemogenetic activation used DREADDs and clozapine-N-oxide (0.5 mg kg⁻¹). During the light phase, 14 mice spent more time in slow-wave and REM sleep, less time awake, and reached sleep sooner (latency P = 0.003). They had more slow-wave-sleep bouts (P = 0.036), while mean bout duration did not differ significantly (P = 0.110). Control mice without the DREADD did not show the same change. A dark-phase experiment comprised 12 sessions from six mice and also increased slow-wave sleep and shortened latency.

4. Reading the numbers

The fractions below are transparent recalculations from reported counts, not new biological measurements. For a count k among n cells, p = k/n. Thus 95/111 = 0.856, or 85.6%, and 14/16 = 0.875, or 87.5%. The smaller deep-layer sample has much wider sampling uncertainty, so the percentages should not be read as precise population rates.

$p = \frac{k}{n}$

$\frac{95}{111}=0.856\; (85.6\%)$

$\frac{14}{16}=0.875\; (87.5\%)$

Measurement	Reported basis	Bounded interpretation
Low-arousal superficial cells	95/111 cells; 15 mice	85.6% of labelled cells in this sample
Low-arousal deep cells	14/16 cells; 5 mice	87.5%; small sample
Slice recordings	156 cells; 16 mice	Long-range GABAergic responses, including ~2 mm
Light-phase DREADD	14 mice; CNO 0.5 mg kg⁻¹	More SWS/REM and shorter latency
Dark-phase DREADD	12 sessions; 6 mice	More SWS and shorter latency

Delta power can be represented conceptually by integrating the spectrum from 1 to 4 Hz; consult the source article for the exact experimental estimator.

$P_{\delta}=\int_{1\,\mathrm{Hz}}^{4\,\mathrm{Hz}} S_{xx}(f)\,df$

5. What the result means—and does not mean

This work shows sufficiency: forced activation can move the system toward synchronization and sleep. It does not show that Sst-Chodl cells are necessary for normal sleep, that the same intervention would work in humans, or that cortex acts alone. The animals were mice; sample sizes differ by experiment; cell targeting is imperfect; and the authors explicitly leave the upstream inputs, roles outside sensory cortex, and connections to hypothalamus or midbrain for future work. The observed association with delta activity was contemporaneous rather than predictively established.

6. Reproducibility and next questions

The paper is open access and provides extensive methods and source data. The final Nature article was published on 9 September 2026. A 2024 bioRxiv preprint documents the earlier version of the same study; it is useful for version history but is not independent corroboration. The most informative next tests are loss-of-function experiments, recordings beyond sensory cortex, mapping of upstream and subcortical partners, and replication in other mammals.

7. Sources

8. News article method

Article written and translated by an AI system from the cited sources, with automated checks under the News editorial method. No human expert review is claimed.