Researchers have developed a new method to study the developing human brain by growing human brain tissue inside the skulls of genetically modified mice. The study, published in Nature, indicates that replacing the mouse’s native cortex with human stem-cell-derived brain tissue allows the human cells to grow substantially, form long-distance connections, and respond to injuries. This approach provides a new way to observe how human brain circuits function and react to diseases in a living organism.
“The main motivation was a fundamental limitation in studying disorders of the human brain,” senior author Sergiu P. Pașca, the Uytengsu Family Founding Director of the Stanford Brain Organogenesis Program at Stanford University, told PsyPost. “Many neurological and psychiatric conditions begin during development, but living human brain tissue is largely inaccessible, and animal models cannot fully reproduce human genetics or aspects of biology that may be specific to our species.”
To bypass this limitation, scientists often grow human neural organoids in the laboratory. These organoids are three-dimensional spheres of brain tissue cultivated from human stem cells. “Over the years, organoids and assembloids have given us increasingly powerful ways to study human brain development in the laboratory,” Pașca explained. “But they still lack many features of an intact nervous system, including a normal blood supply, sensory inputs and motor outputs.”
While lab-grown organoids can mimic basic cellular processes, they lack the complex environment of a real body. A 2024 study involving the transplantation of similar lab-grown neural clusters into mice indicated that transplanted human cells can mature and respond to inflammation. “It is also difficult in a dish to connect changes in human neural circuits to functional consequences,” Pașca added.
Pașca and his colleagues sought a way to let these human brain clusters mature in a living animal. “So the question was whether we could create a model in which human cortical tissue could develop more extensively, integrate into an intact nervous system and allow us to study human biology from cells and circuits all the way to functional readouts,” he told PsyPost.
In traditional transplants, however, the human cells must squeeze into whatever physical space is left in the animal’s intact skull. They also have to compete with the host’s existing neural circuits. “The two parallel developing systems are in competition for turf,” Pașca noted. To achieve this, the researchers genetically modified mice to lack a neocortex and hippocampus.
The neocortex is the folded outer layer of the brain involved in higher-level functions like sensory perception and spatial reasoning, while the hippocampus manages memory. By using a specific genetic targeting strategy, the team prompted the cells destined to become the cortex to die off early in the mouse’s development.
The resulting animals, termed apallial mice, lose about half of their total brain volume. Despite lacking these major brain structures, the mice survive and can walk around relatively normally. “Our findings suggest that, when cortical circuitry is lost very early in development, the cortex may not be solely responsible for all of the functions traditionally attributed to it,” Pașca said. “Other brain regions may compensate for some of the missing cortical circuitry as the brain develops.”
The researchers then took human cortical organoids and transplanted them into the empty brain cavities of newborn apallial mice, referring to this large-scale grafting procedure as xenocortication. The human grafts grew massively over the next few months. By the third month, the transplanted human tissue exhibited a 4.7-fold increase in size, occupying nearly 92 percent of the available cortical space in the mouse’s brain.
The human cells developed into a diverse array of mature neural cell types. Notably, the graft successfully generated layer 5 extratelencephalic projection neurons. In humans, this specific class of neurons includes von Economo neurons, which are large, specialized cells found in apes and humans that are associated with social behavior and certain psychiatric conditions.
Organoids grown in laboratory dishes almost never generate these specific neurons, but the transplanted organoids in the mice produced them in abundance. “This cell type appears to be particularly vulnerable in frontotemporal dementia, a neurodegenerative disorder that can begin in midlife,” Pașca explained. “Yet here they were, sitting in the xenocortical mice’s human tissue. Now we can generate these rare cells from a healthy person and study them in a living, behaving animal to learn more about what they’re doing.”
The researchers also observed extensive physical integration between the human tissue and the mouse host. Using fluorescent tracing techniques, they saw human nerve fibers extending deep into the mouse’s brain and traveling all the way down into the host’s cervical spinal cord. In return, the mouse’s lower brain regions sent connection fibers up into the human graft.
To test if the human brain tissue was actually functioning, the team used advanced imaging and electrical recording techniques. They detected synchronized bursts of electrical activity spreading across the human graft, resembling the organized brain waves seen in early human development. This spontaneous neural activity strongly correlated with the physical movements of the mouse’s face, suggesting the human tissue was electrically active and participating in the animal’s nervous system.
Next, the researchers observed the animals’ behavior using a combination of motion-tracking cameras and standard laboratory tasks. Both the completely apallial mice and the mice with human grafts walked at normal speeds and explored their environments. However, the xenocortical mice displayed unique behavioral patterns that sat somewhere between the fully cortex-less mice and normal, healthy mice.
In a working memory task using a Y-shaped maze, normal mice tend to remember which arms of the maze they have already explored and alternate their choices. The apallial mice failed to alternate above chance levels, suggesting a loss of working memory. Interestingly, the mice with the human grafts performed above chance, showing a partial retention of this exploratory memory function.
Finally, the team tested whether the xenocortical mice could be used to model human brain injuries. They exposed the animals to a low-oxygen environment (5 percent oxygen) for five hours to simulate a hypoxic injury, a condition that can cause cerebral palsy and motor deficits in human infants.
A 2012 study evaluating gait in mice with Parkinson’s-like symptoms demonstrated that an automated glass walkway system, known as CatWalk, can precisely measure footprint patterns, limb support, and timing. Using this exact system, the researchers evaluated the walking patterns of the xenocortical mice before and after the low-oxygen injury.
Following the injury, the human grafts showed increased signs of cellular stress and inflammation. Behaviorally, the xenocortical mice altered their walking patterns on the glass walkway. They spent more time supporting their weight on three or four paws at once, adopting a wider stance compared to their pre-injury walking style, which provides evidence that the xenocortical mouse model can successfully translate a cellular injury into a measurable behavioral change.
“Finding out what accounts for this difference could yield clues about human neural susceptibility to oxygen deprivation, shed light on mechanisms underlying cerebral palsy and provide a platform for testing potential therapeutic strategies,” Pașca noted. Furthermore, because researchers can derive organoids from specific individuals, the model opens new doors. “The cells we implant carry the genetic material of the person they’re derived from — whether that person is a patient or a healthy individual — allowing us to study downstream disease effects in brain cells and circuits,” he said.
“The idea that you can make an organoid model with an individual’s unique genetic character and use that to learn what’s gone awry in that individual’s brain is a critical step toward precision medicine,” added Alison Singer, president of the Autism Science Foundation.
As with all research, there are a few things to keep in mind. One primary limitation is the biological mismatch in developmental timing. Mouse brains mature over a matter of weeks, while human brain tissue requires years to fully develop. This discrepancy in speed might restrict how completely the human cells can integrate into the rapid life cycle of a rodent.
It is also unclear exactly how the human graft influences the mouse’s behavior. The observed behaviors are likely a mix of the human graft interacting with the mouse’s preserved lower brain circuits rather than the human tissue exerting complete, top-down control over the animal.
“This is not a ‘human brain in a mouse’,” Pașca told PsyPost. “It is a research model that may allow us to understand why human neural cells and circuits become vulnerable in disease and, ultimately, to test ways of preventing or correcting those changes.”
Generating these advanced models also requires careful ethical oversight. “Throughout several years of experimentation, we have received input from ethicists, neurobiologists with expertise in primate and human cortical biology, patient advocates, philosophers, and legal scholars,” Pașca said. “An overriding argument questioned the ethics of not conducting this research in the face of the suffering of hundreds of millions of people afflicted with neurological disorders that today are uncurable but tomorrow could yield treatments we discover by using this model.”
The study, “Developmental xenocortication using human-derived organoids in mice,” was authored by Konstantin Kaganovsky, Kevin W. Kelley, Tilo Gschwind, Paul M. Harary, John Kochalka, Alexander D. White, Garikoitz Lerma-Usabiaga, Xiaoyu Chen, Omer Revah, Felicity Gore, Ayano Aoyama, Jennifer L. Shadrach, Se-Jin Yoon, Alfredo Valencia, Satoe Ogawa, Noah Reis, Hannes Vogel, Brian Wandell, Julia A. Kaltschmidt, Ivan Soltesz, Karl Deisseroth, and Sergiu P. Pașca.