A link appeared in the morning Hacker News digest to a Quanta Magazine article with phrasing that no biologist—and certainly no engineer—could pass up: ctenophore nerve cells are connected by continuous cytoplasm, without a single synapse, and "no other animal has anything like it." The hook wasn't just "another wonder of nature"—it was an architectural break: a nervous system that violates the main rule of neurobiology, formulated by Ramón y Cajal himself, has been discovered in the most ancient branch of the animal tree. That is, evolution apparently invented neurons not once—it did so at least twice, and the second time—completely differently from ours.
I dove into the primary sources and fell down a rabbit hole for a week of scientific papers: chromosomal syntenies, volume electron microscopy, chromatin loops, invasion of the Black Sea, economic damage in the hundreds of millions of dollars. The story turned out to be far sweeter than the announcement. Below is its full breakdown.
Comb jellies, or ctenophores (Ctenophora), are about two hundred species of gelatinous marine animals, from deep-sea giants to coastal translucent "bowls." They swim using rows of comb plates—thousands of fused cilia—and it's precisely these rows that produce that famous iridescence: rainbow waves run along the body not because the animal glows, but because the ciliary bundles work like diffraction gratings and decompose white light into the spectrum. Divers describe them as "barely visible translucent ghosts with iridescent combs." They can actually glow for real—most species have bioluminescence, and the same family also produces coelenterazine, the ocean's most widespread luminescent compound: it's found in copepods, shrimp, squid, and mollusks.
Externally, ctenophores were confused with jellyfish for centuries, and in English they're still called comb jellies, though they have nothing to do with jellyfish. But the kinship here is fictitious: at the genome level they're closer to sponges, which have no nervous system at all. And here's where the detective story begins.
The school textbook long said simply: first came sponges—primitive, without nerves and muscles, then everything else. The order of branch separation on the phylogenetic tree has direct bearing on what the first animal was like: the branch that separated first is the sister group to all others, and its anatomy tells us more about our common ancestor than anyone else.
The problem is that ancient animals' soft bodies don't leave fossils—you have to reconstruct the tree from living species, and morphology and genetics stubbornly gave different answers. The turning point came in 2008, when Mark Martindale, Gonzalo Giribet, and coauthors analyzed genes from 29 species and proposed a radical version: the first to separate were not sponges, but ctenophores. "The entire scientific community ridiculed us," Martindale recalled.
The final move was made by Schultz and coauthors in Nature, May 2023—using a method worth describing separately. Instead of comparing gene sequences (the data have been contradictory for decades), they compared the order of genes on chromosomes—synteny. Genes rarely move from chromosome to chromosome, and such events are almost irreversible. It turned out that sponges, bilaterians (including us), cnidarians, and placozoans carry common traces of rare fusions and reshufflings on their chromosomes, while ctenophores do not. Ctenophores group with unicellular animal relatives by ancient chromosomal patterns, while all other animals group with each other. Conclusion: ctenophores are the sister group to all other living animals. The authors call the support "robust and unambiguous"—a rare formulation for phylogenetics.
Hence the paradox, formulated by Paul Burkhardt himself from the University of Bergen: "If you think about the earliest branch of the animal tree, you expect less complexity." But ctenophores have muscles, nerves, and complex behavior, while sponges have none of this. Either the first animal ancestor was more complex than we thought, or the nervous system appeared in ctenophores separately from ours.
Here we need to step back a century and a half. In the 1880s a famous dispute raged: Camillo Golgi and Joseph von Gerlach believed the nervous system was a continuous network, within which cytoplasm flowed uninterrupted; Santiago Ramón y Cajal defended the "neuron doctrine"—the nervous system consists of separate cells connected by contacts. They worked with the same Golgi stain; both were given the Nobel Prize in 1906—one of the most ironic ceremonies in the history of science, the laureates never acknowledged each other's correctness. The electron microscope in the 1950s closed the dispute: the synaptic cleft—the gap between neuron membranes—is real, Cajal is right, case closed.
Or so it seemed. In Science, April 2023, Burkhardt's lab published a three-dimensional reconstruction of the subepithelial nerve net of the young cydippid stage of the ctenophore Mnemiopsis leidyi, imaged using serial block-face scanning electron microscopy (SBFSEM) with high-pressure freezing—a technique that preserves ultrastructure with almost no fixation artifacts. Result: the nerve net neurons are one syncytium. The plasma membrane between them is uninterrupted: their neurites fuse, forming a single continuous network. There are no typical presynaptic triads between neurons. There are also no electrical synapses (gap junctions)—innexins, the proteins from which they're built in invertebrates, aren't expressed in these neuron bodies at all. At the same time, DiI dye injected into one neuron freely spreads throughout the entire network—physically it's one whole.
That is, Golgi was wrong about our nerves—but turned out to be right about ctenophores. The reticular theory, buried 70 years ago, turned out to be a description of an alien nervous system. As the authors themselves formulated it: the similarity of ctenophore nerve nets to those of other animals can only be superficial—their connectivity is fundamentally different.
How then is the signal transmitted without synapses? These neurons have ion channels—one voltage-gated calcium, thirty-five potassium, and two nonspecific sodium—so the membrane is capable of generating action potentials. But the authors' main bet is volume transmission: many peptide vesicles along the entire network, from which neuropeptides are released directly into the mesoglea (gelatinous matrix), as in the neuroendocrine system, not targeted into a synaptic cleft. It's also consistent that conductivity in ctenophores is demonstrably slower than in other animals—for broadcast diffusion, speed isn't needed. Synapses, by the way, do exist in ctenophores—but not between network neurons, only at the output: on effector cells, for example on the balancer cilia of the statolith.
The aboral organ is the ctenophore's "roof": a sensory hub through which the animal senses gravity (a statocyst with eight lithocysts on four groups of balancer cells with "sickle-shaped" cilia), light and pressure, and which commands the comb plates. In Science Advances, March 2026, Ferraioli, Digel, Colgren, and colleagues from Burkhardt's lab published a complete three-dimensional cellular atlas of this organ in a one-day-old larva. Five SBFSEM datasets, manual segmentation, and the result: nearly 900 cells divided into 17 types, of which 11 are described for the first time. Secretory cells of three sorts, dome cells, lamellar bodies, "feather-like" cells, polar field cells, large granular cells, vesicle-rich nonsynaptic cells—and all of them integrated with that same syncytial nerve net, which thickens around the organ, forming a "condensation."
The final touch was laid by work published in eLife in February 2026: the complete connectome of the statocyst—1011 cells—showed that syncytial neurons form synapses on balancer cells and reciprocal connections with "bridge" cells, and the network itself works as a coordinator of comb cilia, not as a classical sensorimotor reflex. High-speed filming revealed that balancer cilia beat and stop in coordination, but differently in the sagittal and tentacular planes—reflecting the network's organization.
And here's the subtlest point for evolutionary interpretation. In cnidarians and bilaterians, larval apical (top) organs are homologous—their kinship is confirmed by shared transcription factors: Hox, FoxQ2, Six3/6, and others. Comparison of the ctenophore aboral organ's transcriptional profiles with these networks yielded an almost empty result: MlFoxJ1 is the only gene in common with apical organs of other animals. Everything else—unique molecular pathways. The authors phrase it cautiously but unambiguously: the ctenophore aboral organ "probably evolved independently in the ctenophore lineage"—by unique composition of cell types, by special neuronal integration, and by absence of conserved gene expression. The same engineering result—a sensory central hub—obtained from different parts.
While marine biologists were cutting larvae into sections, genomicists were asking a parallel question: where did complexity come from at all? After all, the genes found in ctenophores and cnidarians are largely the same as in their unicellular predecessors. "At the beginning of animals there wasn't so much innovation in genes," says Arnau Sebé-Pedrós from the Centre for Genomic Regulation in Barcelona. "What was important was the ability to modularly regulate these genes in different combinations in space and time."
The mechanism that provides this is three-dimensional chromatin packaging: DNA sections extend into loops that bring distant enhancers close to promoters, and loops assemble into topologically associating domains, TADs—genome "neighborhoods." In Nature, May 2025, Kim and Sebé-Pedrós's team with the Micro-C method went through the three-dimensional genome structures of cnidarians, ctenophores, and placozoans—and unicellular relatives. Unicellular organisms have no loops. Ctenophores do, and many: more than four thousand loops in a genome of only 100 million base pairs (for comparison: humans have 3 billion). Consequently, loop regulation arose ~150 million years earlier than thought—and was probably the lever that allowed the same set of genes to assemble different tissues: nerve, muscle, sensory. Curious detail: ctenophores make loops without the key protein CTCF, which in us holds loop boundaries—probably another protein from the same family does the same work. It's like finding someone else's code solving the same problem with a different API.
Ctenophores also have a trick that broke another dogma. Reverse development—the ability of an adult to return to an earlier stage of the life cycle—was considered exclusive to cnidarians, and the "immortal jellyfish" Turritopsis dohrnii was the only species capable of this after the start of sexual reproduction. In PNAS, 2024, Soto-Angel and colleagues showed: under stress (starvation) and subsequent feeding, the adult lobate form of M. leidyi completely rolls back to the cydippid stage—its lobes and auricles disappear, tentacles regrow, and all 20 tracked individuals went through this path. It's not immortality, but a demonstration that in the most ancient branch of animals the developmental program isn't a one-way arrow, but a two-way conveyor that can be rolled back. For understanding how animal life cycles are actually structured—and ours, the human one, can't be rolled back—this is a fundamental shift.
Every model organism's history has a human chapter. For Mnemiopsis leidyi it's grim. In the early 1980s (according to US State Department data—in 1982) the ctenophore entered the Black Sea with ballast water from ships from the Atlantic coast of America. In the new environment without enemies, being a hermaphrodite with self-fertilization, it staged a demographic explosion: the biomass peak came in autumn 1989, and by some estimates of that period it constituted more than 90% of the entire Black Sea biomass. Zooplankton—its food—was eaten up; then ichthyoplankton (fish eggs and larvae) crashed; the Azov anchovy fishery, already undermined by overfishing, simply collapsed. Estimates of economic damage for the first 18 years—over 350 million dollars (Bright, 1998; estimate cited by the US State Department). From the Black Sea, mnemiopsis seeped into the Marmara (1989–1990), into the Azov (1988), and into the Eastern Mediterranean (1993).
And then what biologists call biological control happened: in 1997–1999 another ctenophore—Beroe ovata, which feeds on mnemiopsis and almost nothing else—arrived in the Black Sea (and later the Caspian) with ballast water. The imbalance was partially compensated, zooplankton and fish stocks began to recover. The moral of this part of the story is unpleasantly honest: the animal whose nervous system we dissect as a marvel of evolution is simultaneously one of the most successful invasive species on the planet, and its inventiveness works both ways.
Let's bring the lines together. Phylogenomics (2023, Nature) places ctenophores at the base of the tree. Superstructural microscopy (2023, Science; 2026, Science Advances and eLife) shows their nervous system is built according to a scheme we thought was disproven 70 years ago—and the molecular profile of this construction is different: ctenophores lack familiar neurotransmitters like serotonin, dopamine, and acetylcholine (glutamate is present), but have their own arsenal of peptides, as Leonid Moroz formulated—"not only different grammar, but a different alphabet." Regulatory genomics (2025, Nature) shows that 3D genome packaging tools appeared 150 million years earlier than thought—that is, the "software" of complex animals was ready before the animals themselves. And developmental biology (2024, PNAS) adds: even the direction of their ontogeny is not law, but a setting.
The most honest conclusion that follows: the question "were neurons invented twice" is not definitively closed. There are two scenarios: either the nervous system arose in ctenophores separately (and our common ancestor was a neural void, and sponges simply never lost anything), or it was in the common ancestor, but in sponges—disappeared completely, which would be no less extravagant: we know complete nervous system loss in bilaterians only in parasites, as Christopher Lowe from Stanford noted. Both variants are about evolution of major solutions not being monopolized by one outcome: convergent origin of eyes in different lineages has long surprised no one, but convergent origin of neurons—that's no longer about form, but about the most complex cell type in nature. "A single neuron is like the sound of one hand clapping," Martindale said. If two different trees independently learned to clap—it means evolution has not only repeatability of solutions, but repeatability of inventing their carriers.
And a practical tail. Placentogens—cone lipids that deep-sea ctenophores use to maintain membranes under thousands of meters of pressure—are also in the human brain; their deficiency is linked to neurodegenerative diseases. Ctenophore peptides are a potentially new pharmacology arsenal. And syncytial architecture is a direct question for synthetic neurobiology: what if a network without targeted synaptic connection isn't a deformity, but an alternative design worth exploring on chips?