---
title: "The Brain’s Two Ancient Origins: What a Landmark Developmental Study Actually Found"
slug: brain-two-ancient-origins-forebrain-hindbrain-2026
category: health
category_label: "Health"
author: "BrainWavePost Staff"
date: 2026-09-23
tags: ["brain development", "neuroscience", "forebrain", "hindbrain", "evolution", "stem cells", "Nature Neuroscience", "Stanford Medicine"]
read_time_minutes: 12
canonical_url: https://brainwavepost.com/article/brain-two-ancient-origins-forebrain-hindbrain-2026
source: BrainWavePost
---

# The Brain’s Two Ancient Origins: What a Landmark Developmental Study Actually Found

*Health · 2026-09-23 · BrainWavePost Staff · 12 min read*

> Mouse-embryo tracing and human stem-cell experiments reveal two parallel progenitor populations for the forebrain–midbrain and hindbrain. The cellular result is strong; the story of two ancient nervous systems joining 550 million years ago remains an evolutionary hypothesis.

> **The careful conclusion** _(note)_
>
> The study directly shows two distinct early neural progenitor populations: one contributes to forebrain and midbrain, while the other contributes to hindbrain. It does not directly demonstrate two adult brains, measure reasoning or consciousness, or date a merger of nervous systems. The proposed 550-million-year history is an evolutionary interpretation that the authors say may explain the developmental split. [1][2]

## A textbook model meets an unexpected split

The vertebrate brain looks continuous, and a longstanding developmental model treated it as the product of one early neural-ectoderm population that was later divided into regional identities. A Nature Neuroscience study led by Stanford developmental biologist Kyle Loh instead reports that two lineage-restricted populations emerge in parallel during gastrulation, the stage when an embryo establishes its major body axes and germ layers. [1][2]

The anterior population contributes to the forebrain and midbrain. The posterior population contributes to the hindbrain. They differ not only in location and marker genes, but also in which parts of their DNA are accessible and in the neural identities they can produce under the tested conditions. That makes the finding more substantial than a temporary difference in gene expression: it points to two early developmental programs with restricted destinations. [1][3]

The authors go further, proposing that this division may preserve the legacy of two ancient neural systems that were integrated during animal evolution. Their paper says the dual progenitors may be conserved across roughly 550 million years, from hemichordates to mammals. That is a biologically informed hypothesis—not a date established by a fossil, a molecular clock analysis or a new experiment on a 550-million-year-old organism. [1][3][4]

- **2** — Parallel neural progenitor populations identified [1]
- **Otx2 / Gbx2** — Key markers separating anterior and posterior programs [1][2]
- **~550 Myr** — Proposed span of evolutionary conservation—not a directly measured merger date [1][3]

## How the researchers followed the cells

The central in-vivo evidence came from developing mouse embryos. The researchers used lineage tracing to mark early cells and follow their descendants, alongside gene-expression measurements during gastrulation. Cells associated with the anterior neural ectoderm expressed Otx2 and contributed to forebrain and midbrain structures; cells associated with posterior neural ectoderm expressed Gbx2 and contributed to the hindbrain. The two marked populations arose at about the same developmental period rather than one appearing as a later branch of the other. [1][2]

The team also compared chromatin accessibility—the pattern of DNA regions open to the molecular machinery that controls gene activity. Anterior and posterior progenitors had different accessible regulatory landscapes. In practical terms, the cells were already reading different sets of developmental instructions before mature brain regions existed. [1][3]

To examine whether the distinction extended to human biology, the researchers directed human pluripotent stem cells along anterior or posterior neural paths in culture. Different signaling conditions generated OTX2-positive anterior-neural-ectoderm-like cells or GBX2-positive posterior-neural-ectoderm-like cells. Mixing and differentiation experiments supported the conclusion that these populations were lineage-restricted rather than interchangeable versions of one generic brain progenitor. [1][2][3]

## What Otx2 and Gbx2 tell researchers

Otx2 and Gbx2 encode transcription factors: proteins that help control networks of genes. In vertebrate development, their expression domains meet near the future midbrain–hindbrain boundary. The new work uses that familiar boundary in a deeper argument: the distinction is present in the progenitors themselves, not merely imposed later on one uniform sheet of neural tissue. [1][5]

A marker alone would not establish ancestry or fate. The case is stronger because the study combines marker expression, lineage tracing, chromatin state and controlled differentiation. Together, those methods connect molecular identity to developmental destination. They still do not make the two populations independent organs inside a modern embryo; the populations develop next to one another and ultimately form a connected central nervous system. [1]

## The most immediate advance: making hindbrain neurons

The developmental split was useful as well as descriptive. By first generating posterior neural ectoderm under conditions suited to that lineage, the researchers produced human motor neurons with characteristics of hindbrain rhombomeres 5 and 6. They assessed the cells with molecular markers and electrophysiology, which tests whether neurons can generate appropriate electrical activity. [1][2]

That matters because protocols that efficiently generate forebrain or spinal-cord identities do not automatically produce authentic hindbrain cells. A clearer recipe could improve laboratory models of development and of conditions affecting hindbrain motor neurons. Stanford’s account highlights amyotrophic lateral sclerosis and spinal muscular atrophy as possible research applications, but the study did not model either disease, test a treatment or show that laboratory-grown cells can repair a patient’s nervous system. [2]

## Where reasoning, consciousness and breathing fit—and where they do not

The forebrain includes structures essential to perception, planning, language and many forms of conscious experience. The hindbrain contains circuits involved in breathing, cardiovascular control, sleep, balance and movement. Those are established anatomical and physiological descriptions that help readers understand why the developmental boundary is important. [2][6]

They were not outcomes measured in this experiment. The researchers did not test reasoning, consciousness, poetry, mathematics, breathing or heartbeat. Nor does an anterior developmental origin make every forebrain function uniquely human: other vertebrates possess forebrains, and many cognitive capacities emerge from interactions among forebrain, midbrain, hindbrain, cerebellar and bodily systems. Turning the paper into a contrast between a ‘thinking brain’ and an ‘automatic brain’ is memorable but biologically too simple. [1][2]

## Why scientists invoke two ancient nervous systems

Hemichordates, including acorn worms, are close invertebrate relatives of chordates. They do not have a vertebrate brain, but they possess patterned nervous tissue and developmental genes that allow researchers to compare how anterior and posterior neural territories may have been organized before vertebrates arose. Comparative work has shown both conserved and lineage-specific features in hemichordate nervous-system development. [4][7]

The authors interpret the anterior–posterior progenitor split in mice and human cells against that broader evolutionary literature. One possibility is that ancestral animals had spatially distinct neural centers that were later incorporated into the continuous head nervous system of vertebrates. If correct, modern embryonic development retains a molecular record of that history. [1][3][4]

But developmental similarity is not a preserved video of evolution. Genes and regulatory programs can be retained, repurposed or independently modified. Demonstrating that two populations exist today does not by itself prove that they once belonged to fully independent nervous systems, exactly when they came together or what the ancestral structures did. Testing the idea will require broader comparisons across living animal lineages and explicit evolutionary analyses. [1][4][7]

## Accuracy check: claim versus evidence

- Supported: forebrain and midbrain arise from an Otx2-associated anterior neural progenitor population, while hindbrain arises from a distinct Gbx2-associated posterior population in the systems tested. [1]
- Supported: the two populations emerge in parallel during mouse gastrulation and have distinct gene-regulatory and chromatin-accessibility states. [1][3]
- Supported in a laboratory model: human pluripotent stem cells can be directed into corresponding anterior- and posterior-neural-ectoderm-like populations. [1][2]
- Supported: recognizing the posterior pathway enabled production of human hindbrain rhombomere 5/6 motor neurons in culture. [1]
- Too broad: ‘the human brain is two separate organs.’ The direct evidence concerns early progenitor lineages; the mature brain is anatomically and functionally integrated.
- Not tested: reasoning, consciousness, language, breathing or heartbeat. References to those functions are explanatory context, not experimental results. [1][2]
- Hypothesis, not direct demonstration: the lineages descend from two independently evolved nervous systems that joined around 550 million years ago. The paper says they may be conserved across that span. [1][3][4]
- Not a clinical result: no patient treatment, transplant, diagnostic test or disease cure was evaluated. [1][2]

## Important limitations

1. Mouse embryos provide the in-vivo lineage evidence. The human portion uses pluripotent stem cells in culture, not lineage tracing in living human embryos. [1]
2. A dish model simplifies the embryo’s geometry, timing and interactions with surrounding tissues. Successful differentiation does not prove that every step perfectly reproduces human development.
3. The evolutionary proposal draws on comparisons with prior hemichordate and animal-development research; this paper did not perform new experiments across 550 million years of lineages. [1][3][4]
4. Otx2 and Gbx2 identify broad developmental territories, but mature brain regions contain many cell types and later subdivisions that this two-part account does not capture.
5. Generating hindbrain motor neurons is a research-platform advance. Their suitability for modeling particular diseases, predicting drug effects or supporting therapy requires separate validation.

## What should happen next

The evolutionary model can be tested more rigorously by comparing matched single-cell gene-expression and chromatin maps across vertebrates, hemichordates and other deuterostomes, then reconstructing when specific regulatory programs appeared. Experiments that perturb the anterior and posterior programs across species could reveal whether similar genes perform genuinely homologous developmental jobs. [4][7]

For biomedical research, the immediate priority is reproducibility: laboratories need to confirm that the posterior protocol consistently generates well-defined hindbrain subtypes, that those cells mature appropriately and that disease-associated changes resemble what happens in people. The study opens that path. It does not complete it. [1][2]

The enduring result is therefore both more precise and more useful than the ‘two brains’ headline. The earliest vertebrate brain does not appear to begin as one completely uniform population. It begins with at least two parallel, molecularly distinct neural programs that become one connected organ. Whether those programs are the living signature of two nervous systems separated by deep time is now a compelling hypothesis for comparative biology to test. [1][3][4]

## Sources

- [1] Nature Neuroscience — Jokhai, Dundes et al., ‘Two parallel neural ectoderm progenitors contribute to the developing brain’ (18 September 2026), DOI 10.1038/s41593-026-02433-7: https://www.nature.com/articles/s41593-026-02433-7
- [2] Stanford Medicine — ‘Human brain is two separate organs, Stanford Medicine-led research finds’ (18 September 2026): https://med.stanford.edu/news/all-news/2026/09/two-separate-brains.html
- [3] bioRxiv — Dundes, Jokhai et al., preprint ‘Two parallel lineage-committed progenitors contribute to the developing brain’ (2025), DOI 10.1101/2025.07.02.662771: https://www.biorxiv.org/content/10.1101/2025.07.02.662771v1
- [4] PLOS Biology — Formery et al., ‘A whole-body atlas of the gene regulatory dynamics of the developing acorn worm Ptychodera flava’ (2023): https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002242
- [5] Nature Reviews Neuroscience — Wurst and Bally-Cuif, ‘Neural plate patterning: upstream and downstream of the isthmic organizer’ (2001), DOI 10.1038/35053516: https://doi.org/10.1038/35053516
- [6] National Institute of Neurological Disorders and Stroke — ‘Brain Basics: Know Your Brain’: https://www.ninds.nih.gov/health-information/public-education/brain-basics/brain-basics-know-your-brain
- [7] Current Biology — Lowe et al., ‘Dorsoventral patterning in hemichordates: insights into early chordate evolution’ (2006), DOI 10.1016/j.cub.2006.05.055: https://doi.org/10.1016/j.cub.2006.05.055

Editorial note: the lead image is an original conceptual illustration of two developmental domains joining in a vertebrate brain. It is not an embryo scan, lineage-tracing image or anatomical reconstruction from the study. BrainWavePost has no commercial relationship with the authors, Stanford University or the cited publishers.

---

_Canonical article: [https://brainwavepost.com/article/brain-two-ancient-origins-forebrain-hindbrain-2026](https://brainwavepost.com/article/brain-two-ancient-origins-forebrain-hindbrain-2026) — © BrainWavePost. Educational content; see the article page for full disclaimers._