485 out of 6,827. That is the number of quantitative checks on recorded brain activity that moved in the same direction in every one of 15 awake-versus-anesthetized comparisons, spanning six species and seven anesthetic drugs or drug combinations. Any single check would pass that test by coincidence about once in 16,384 tries — the odds of a coin landing on the same face 15 times running — so chance alone accounts for roughly 0.4 of the 6,827. The 485 that passed describe the same two failures in every animal, George Semaan writes at The Daily Neuron: each patch of brain activity stops echoing its own recent past too quickly, and the patches fall out of step with each other.
The study, by Andrea Luppi and 27 colleagues in Nature Neuroscience, runs from one extreme of experimental neuroscience to the other. In one arm, 15 healthy human volunteers breathed sevoflurane through a mask while the dose stepped up every three minutes; each was asked repeatedly to squeeze a hand, and dosing climbed until two requests in a row went unanswered. In another, 10 worms lay in gel under isoflurane while a microscope tracked about 120 neurons per worm, made fluorescent by a gene that lights up active cells. Functional MRI covered the humans, macaque and marmoset monkeys, and mice; microscopes covered worms and larval zebrafish, the fish recordings being the only data collected new for this study.
The analysis deliberately refuses to guess what matters. Every raw trace — from a single worm neuron to a whole human brain region — was run through 6,827 measures borrowed from fields as distant as physics and economics, each reducing a wiggly line of activity to one number. A measure counted as a shared signature of anesthesia only if it moved the same direction, up or down, in all 15 comparisons. Two obvious confounders — body motion and recording noise — failed that same test, which is the closest the design comes to a built-in control.
An echo that dies early
In a waking brain, what a region does now retains a trace of what it did a moment ago; neuroscientists call the length of that trace the intrinsic timescale. Under anesthesia, in all six species, the trace shortened. Regions with the longest timescales while awake lost the most — and long timescales are exactly what is thought to let a region integrate information over time. The authors’ suggestion is that with shorter timescales, a stimulus, even a painful one, may decay before it propagates far enough to matter. The study’s own recordings show slow activity speeding up under anesthesia, while earlier work found fast brain waves slowing down; the two results together point to a squeeze toward a narrower, less diverse range of timescales.
A band that loses the beat
The second signature is synchrony. Distinct regions of a waking brain tend to rise and fall together; under anesthesia, across species and drugs, that coordination weakened, even between regions whose activity had similar shapes. Loss of coordination between dissimilar regions would be unsurprising — losing it between matched ones suggests the coupling itself, not just the content, is failing.
Awake on a steady drip
The strongest causal evidence comes from two macaques in an earlier experiment by Jordy Tasserie, Bechir Jarraya and colleagues. Both monkeys sat on a steady intravenous flow of propofol — the drug present, at anesthetic concentration, the whole time. Electrical pulses to a small site in the thalamus, the deep-brain structure that helps maintain arousal, woke both animals: they opened their eyes and responded while the drug kept flowing. Pulses of identical strength delivered elsewhere in the thalamus did nothing. Luppi’s team counted a change as a true signature of unconsciousness only if it also reversed during those pulses: when the monkeys woke on propofol, the echoes lengthened again and regions re-synchronized. That design separates the drug’s presence from its effect, which is the confound that runs through most anesthesia neuroscience.
Different chemistry, the same endpoint
The drugs in the study do not share a target. Propofol and the inhaled anesthetics prolong inhibitory signalling — the chemistry that tells the next neuron not to fire. Ketamine blocks a major excitatory signal instead. Either way, a neuron has a harder time summing its inputs into a response, and the shared macroscale pattern emerges anyway. In the four mammals, the spatial pattern of change aligned with expression maps of 23 genes: inhibitory signalling genes at one end, excitatory genes at the other, including genes for the molecular targets propofol and ketamine bind to — a hint that the four mammals share the machinery the drugs act through.
A computational model of the human brain, taken unchanged from earlier work and set to a propofol-like regime — longer inhibition, faster neuronal fatigue — reproduced 461 of the 485 shared changes, with shortened timescales included. A ketamine-like setting reproduced more than 90 percent. The model is offered as one candidate path from synaptic-scale changes to the brain-wide pattern, not as proof of mechanism.
What it does not settle
The result lands on a two-decade-old idea: Max Kelz and George Mashour argued in 2019, reviewing evidence from paramecia to primates, that every effective anesthetic leaves an organism disconnected from the world and unable to exchange information internally. It also converges with Laura Lewis’s 2012 electrode recordings from three epilepsy patients going under on propofol. And the same shortened timescales appear in deep surgical anesthesia and in patients with disorders of consciousness — while light, sedating doses stretch the timescale instead. The caveats are the study’s own: unconsciousness was judged by behaviour (a worm cannot squeeze a hand), the work says nothing about what any animal experiences, and clear differences between species and drugs remain inside the shared pattern.
By the authors’ count, 700 million years of evolution separate the worm from the volunteer with the mask. The practical proposal is a yardstick: because the 6,827 measures run on any recording from any nervous system, the authors suggest they could serve as a common, cross-species readout of how far under a brain is. Whether any hospital ever monitors a patient that way is the open question — the next number to watch is whether the signature tracks depth of anesthesia in real time, in an operating room, rather than in a retrospective analysis.