25 SEP 2026 — The standard model of brain development starts with one progenitor cell giving rise to the whole brain. A Stanford team reports that there are two progenitor populations instead, that they separate before there is a brain to speak of, and that neither can become the other.
The same division turns up in acorn worms, which last shared an ancestor with us around 550 million years ago.
What the old model said
The standard account has the brain descending from a single population of neural progenitors, with forebrain, midbrain and hindbrain branching off a shared lineage as development proceeds. A branching tree implies a common trunk. If a common trunk exists, one region's cells could in principle be coaxed into another's.
The paper, published in Nature Neuroscience on 18 September, describes no trunk at all.
Two lineages that never cross
Two progenitor populations appear during gastrulation, before the neural tube forms. One expresses Otx2 and is committed to the forebrain and midbrain. The other expresses Gbx2 and builds the hindbrain, the part that runs breathing, heart rate and swallowing.
They do not overlap, and one does not convert into the other. The hindbrain lineage runs alongside the forebrain lineage rather than branching from it.
Kyle Loh, associate professor of developmental biology at Stanford Medicine, is the senior author. The co-first authors are graduate students Carolyn Dundes and Rayyan Jokhai.
The chromatin is packed differently
Epigenomics provides the evidence that the separation is permanent, not merely early. The two populations organise their DNA differently; chromatin architecture differs between anterior and posterior neural ectoderm in ways that lock each cell into its path.
That is a physical answer to a developmental question. The cells do not decline to change identity; the packaging has already closed the door.
It holds across 550 million years
The two-origin pattern was found in mice, chickens, zebrafish and acorn worms, a marine invertebrate whose lineage diverged from ours roughly 550 million years ago. Jellyfish, further back still, have two separate nervous systems outright.
Conservation across that span argues against a mammalian quirk. The vertebrate brain looks assembled from two nervous systems that were already separate, not elaborated from one.
Why hindbrain neurons would not grow
The finding had an immediate practical consequence: the team produced functional human hindbrain motor neurons from pluripotent stem cells. These included cells expressing markers for facial musculature and for swallowing, which had previously been difficult to generate in the laboratory.
Jokhai explains why earlier attempts failed: "Previous attempts to make hindbrain neurons likely tried to coax forebrain and midbrain progenitors into hindbrain cells, which our study shows is not possible."
Laboratories had been starting from the wrong cell. Hindbrain motor neurons are lost in amyotrophic lateral sclerosis and spinal muscular atrophy, so a reliable method for making them opens a route to studying both diseases.
View author profile → · Editorial policy
