Two ancient developmental programs appear to build the human brain
A basic assumption in neuroscience just broke open. Stanford Medicine researchers found that the human brain does not begin as a single uniform structure, but from two distinct progenitor systems that follow separate developmental paths from the earliest stages of embryogenesis. One pathway gives rise to the forebrain and midbrain. The other builds the hindbrain, including the brainstem structures that govern breathing, heartbeat regulation, swallowing, and other automatic life-sustaining functions.
The finding reframes the brain as a product of two ancient neural lineages fused over evolutionary time. It also helps explain a frustrating laboratory problem that has slowed work on ALS, spinal muscular atrophy, and other disorders affecting hindbrain neurons: scientists may have been trying to generate the wrong cell fate from the start.
The hindbrain does not branch off from the forebrain lineage
For years, the prevailing model held that forebrain, midbrain, and hindbrain all emerged from one shared embryonic progenitor population. The new work challenges that view at its foundation. During gastrulation, when the body’s basic blueprint begins to form, the research team identified a developmental split that occurs much earlier than expected. The hindbrain does not appear as a later offshoot of the same pathway that yields the forebrain and midbrain. It develops in parallel.
Two gene programs defined the split. One progenitor population expressed Otx2 and was destined to form the forebrain and midbrain. A separate population expressed Gbx2 and developed into the hindbrain. The populations stayed separate and showed no overlap at the stages examined. Even more revealing, the chromatin landscapes of the anterior neural ectoderm and posterior neural ectoderm were fundamentally different, meaning the cells were already carrying distinct accessibility patterns that locked them into different fates.
A one-line laboratory lesson follows from that biology: if a cell is never poised to become hindbrain tissue, forcing it there is unlikely to work.
Early chromatin state may explain why hindbrain neurons were so hard to grow
That developmental division helps make sense of why human hindbrain neurons have been so difficult to generate in culture. Previous approaches likely tried to push forebrain and midbrain progenitors into a hindbrain identity after they had already crossed a developmental boundary. The investigators argue that this strategy was biologically mismatched to the way the brain is actually assembled.
By focusing on the earliest embryonic time points, the team identified a cellular fork in the road before the lineages fully diverged. This matters because stem cell models depend on recreating the correct sequence of developmental decisions, not just the final cell type. In this case, the developmental sequence appears to matter as much as the endpoint. Without access to the right progenitor state, the desired neuron type may never appear.
Human pluripotent stem cells could finally be directed into hindbrain motor neurons
Once the developmental logic became clear, the researchers used it to solve a long-standing cell culture challenge. Human pluripotent stem cells were successfully guided into functional hindbrain motor neurons for the first time in this study. These cells generated action potentials and produced proteins associated with hindbrain regions that control facial and swallowing muscles.
That achievement is more than a technical milestone. It creates a usable human model for neuronal populations that are nearly impossible to obtain directly from patients. Brainstem tissue cannot be sampled routinely in living people, yet these neurons are central to some of the most disabling neurological diseases. A dish-based system that faithfully reproduces hindbrain biology offers a path to study disease mechanisms, test candidate therapies, and examine why specific motor circuits fail.
The implications are particularly strong for conditions in which bulbar and brainstem dysfunction drive morbidity. In spinal muscular atrophy, a leading genetic cause of death in children under 1 year of age, vulnerable motor neurons progressively fail. In amyotrophic lateral sclerosis, usually diagnosed between ages 40 and 70, hindbrain and forebrain circuits can both be affected. As swallowing weakens, food or liquid can enter the lungs and trigger pneumonia; later, respiratory failure becomes the major threat.
An evolutionary split older than vertebrates themselves
The developmental divide did not appear to be a human quirk. Looking across more than 550 million years of evolution, the researchers found a similar two-origin arrangement in chickens, zebrafish, and even acorn worms, which share a distant common ancestor with humans. Jellyfish, which diverged from the human lineage roughly 600 to 700 million years ago, also possess two nervous systems positioned at opposite ends of the body.
Those comparisons suggest an evolutionary story in which separate neural systems existed first and were later brought into closer physical proximity in vertebrates. Rather than inventing one unified organ from scratch, evolution appears to have repurposed and fused existing neural architectures. The result is the modern brain: anatomically integrated, but developmentally split.
Brainstem disease research may get a sharper experimental model
The practical value of this work lies in disease modeling. Hindbrain neurons are not just one neuronal class among many; they sit at the center of respiratory control, swallowing, facial motor function, and other circuits that are frequently disrupted in neurodegenerative disease. Researchers studying ALS and SMA have been constrained by limited access to living human tissue and by incomplete lab models that failed to generate the relevant neurons with fidelity.
A human hindbrain system grown from the correct progenitor lineage should make it easier to define what goes wrong when these neurons fail, whether the problem lies in early developmental programming, maintenance of neuronal identity, electrical function, or downstream degeneration. The model may also help clarify the developmental origins of the spinal cord, which the team plans to examine next.
The study stops short of answering every translational question. It does not establish a therapy, and it does not yet show that the new cells can reproduce every aspect of human disease. It does, however, redraw the developmental map in a way that changes where the field will look next. For brainstem disorders, that may be the most consequential shift of all.