Last week's Nature paper from Sergiu Pașca's lab at Stanford University, US, led by Konstantin Kaganovsky, is one of the most striking pieces of neuroscience I have read this year. Sergiu and I discussed the future of animal use in brain research at The Brain Debate at the FENS Forum in Barcelona in July. It is also news that those who believe we are on the threshold of ending animal use will not want to hear. 

Human neurones are large, slow-growing, and have nowhere to go in a mouse skull. So, the team used a mutant mouse that never develops most of its cortical and hippocampal structures. Where the mouse's own neurones failed, the human organoid flourished: after three months, more than 90% of the cortical tissue by volume was human, integrated into the host's circuits and supporting normal behaviour. The grafts even produced Von Economo neurones — large, distinctive cells implicated in frontotemporal dementia that nobody has managed to generate in a dish. 

So what can this actually tell us about disease? In reality, much more than any dish-based model could. Yes, this is only a proof-of-concept experiment, but the team exposed xenocortical mice to five hours of severely low oxygen. Oxygen deprivation during pregnancy or around birth is a known cause of cerebral palsy, which affects 3 in 1,000 people and is itself a risk factor for epilepsy and autism. The human cortical tissue was substantially damaged, and the animals went on to have trouble sustaining a steady gait and keeping their balance — reminiscent of what is seen in children with cerebral palsy — while normal mice and cortex-depleted mice without a human graft were virtually unaffected. That contrast is the crux: it isolates a vulnerability specific to human neurones, in a living body, with a behavioural result you simply cannot get from tissue in a dish.  

As Pașca notes, finding out what accounts for that difference could shed light on the mechanisms underlying cerebral palsy and provide a platform for testing therapeutic strategies. The same logic extends: build the organoids from cells donated by patients carrying disease-associated genetic changes and the method could in principle be extended to schizophrenia, epilepsy, autism and some forms of dementia. 

None of this is a finished tool. A real cortex is built in distinct layers, and it depends on cells that put the brakes on activity as well as cells that drive it. The human tissue in these mice has neither: it does not form those layers, and the braking cells never appear. Pașca himself is clear that this is not a universal system for every question about the brain. 

These are not marginal human conditions — more than 1 in 100 adults has schizophrenia, and some 1% of people remain epileptic throughout adulthood — and they are brutally hard to study because living human brain tissue is almost neveraccessible. Pașca's summary is worth quoting: “While animal models have been extremely helpful, some biological features seem to be uniquely human.” He does not say the animal has been superseded. He says human biology had to be put inside one. 

But consider why anyone would do this at all. Rodents remain our most important models of brain function in health and disease, yet a mouse brain is not a human brain and not everything translates. That gap is exactly what human organoids were meant to fill, and exactly why campaigners and activists have spent years insisting organoids will make animal research obsolete. However, organoid researchers themselves will tell you otherwise. These are not brains. They receive no sensory input from a body and send no output to one; they have no blood supply and no immune system. Which leaves an obvious conclusion: to make organoids behave more like brains, you have to put them inside a living animal. This though is only the start. It will take decades to establish how human cells integrate with a host, which functions they acquire, which circuits they form, and what any of it tells us about the human brain. 

On the question some are asking — are these animals in some sense human-minded? No. Mice are sentient. But sentience is not what we mean by a human mind. A mouse does not know that it is a mouse. It has no concept of itself as a being with a past and a future, no reflective awareness of its own thoughts, no capacity to ask what it is. Nothing here changes that. The mice performed much as ordinary mice of the same age did — no new capacities, nothing the host brain could not already do. A mouse-sized graft of unlayered human tissue, in a mouse body with mouse senses and a mouse lifespan, does not produce a human mind. It produces a mouse with a more informative cortex. 

To those who claim we are on course to abandon the use of animals in scientific research, this study is the strongest evidence yet that they could not be more wrong. We will need animals for decades to come — and we now have one more tool to use alongside them. 

EARA Executive Director