Human cortical organoids form working circuits in cortex-depleted mice

The study, published in Nature on 16 September 2026, used a genetic strategy to deplete most glutamatergic neurons from the mouse neocortex and hippocampus. Researchers then transplanted human stem-cell-derived cortical organoids into the resulting cortical cavity shortly after birth. Removing much of the competing mouse cortical tissue gave the human grafts substantially more space to grow and integrate than in earlier transplantation models.

The grafts expanded through much of the available cortical volume and generated a range of human cortical cell types, including layer-5 projection neurons. Calcium imaging and electrophysiology showed organised patterns of activity resembling developing neural circuits, while anatomical tracing showed human neurons connecting with the mouse nervous system and extending projections beyond the graft itself. The work is an experimental animal study and a peer-reviewed version of record, not a clinical trial.

Human brain organoids grown in dishes can reproduce parts of early neural development, but they lack the circulation, sensory input and whole-organism circuitry of a living nervous system. Earlier grafting experiments improved on that limitation but left human tissue competing with an intact host cortex. The new xenocortical model changes that constraint by creating a large developmental niche for the graft and allowing circuit-level and behavioural readouts in the same animal.

The researchers demonstrated one such use by exposing animals to reduced oxygen. Human-derived cortical tissue showed substantial injury, and xenocortical mice developed gait and coordination problems that were not seen to the same degree in ordinary or cortex-depleted control mice. The team presents this as a way to study mechanisms relevant to developmental brain injury and potentially to disorders that are difficult to model using isolated cells alone.

The transplanted tissue remained developmentally immature and did not recreate a normal human cortex. The host mice were also highly engineered animals missing most of the cortical and hippocampal glutamatergic system, so results cannot be directly generalised to an intact human brain or even to an ordinary mouse. Behavioural differences therefore have to be interpreted as properties of this specialised model rather than evidence that the graft reproduces human cognition.

The scale of human-neural integration also raises animal-welfare and neuroethics questions. Stanford says the work underwent institutional review and external ethical consideration, and the reported animals retained broadly preserved locomotion with selective behavioural differences. The study does not establish human-like consciousness, identity or cognition in the mice. Its immediate significance is methodological: it provides a larger, living environment in which human neural tissue can mature, connect and be experimentally perturbed while researchers can observe circuit and behavioural consequences.