How Far Can Human Neural Tissue Compensate for Lost Function?

Human-derived neural tissue is transplanted into mice missing most of their cerebral cortex. A paper in Nature reports how that tissue grows and relates to the animals' behavior. The grafts grew extensively, and connections to the host and neural activity were confirmed. The study offers a way to investigate human neural tissue within a living brain. Kaganovsky et al., Nature 2026 ↗

We want to read this paper by asking which functions transplanted tissue might help compensate for, and what is still missing. Comparisons with studies transplanting dopamine neurons and inhibitory cells offer further clues for interpreting these results.

First, the starting conditions matter. These mice have a substantial reduction in excitatory neurons in the neocortex and hippocampus from early development. The human cortical organoids—immature neural tissue grown from stem cells—were also transplanted when the mice were young. This was not a study in which a fully developed brain was damaged and then restored through transplantation.

What Could the Mice Do, and Where Did They Struggle?

Mice without grafts could walk too. Walking after transplantation is therefore not, in itself, evidence of recovery. However, the time their paws spent off the ground during walking was shorter in the nongrafted group, whereas no difference was detected between the grafted group and controls. Coordination of diagonal paws landing together, on the other hand, remained different from controls after transplantation. Some aspects of individual paw movement and the coordination of multiple paws showed different patterns of results.

Memory-related tasks also produced contrasting results. The Y-maze tests whether mice explore different arms in sequence rather than repeatedly choosing the arm they have just entered. The grafted and control groups showed performance suggestive of exploration using short-term memory; this was not confirmed in the nongrafted group. By contrast, freezing when returned the next day to a place where they had received an electric shock was weak even in grafted mice. Original paper: behavioral experiments ↗

So far, one reading is that there may be scope to compensate for functions supporting exploration over short periods, while difficulties remain in linking a place with an experience and responding to it later. Yet even the encouraging signs in walking and maze performance have not established an improvement over mice without grafts. We would treat these findings not as a list of recovered functions, but as a way to narrow down which functions transplantation might help support.

What questions does this contrast let us ask? It is not necessarily as simple as remembering briefly and forgetting by the next day. The two tasks differ both in what must be remembered and in the behavior they require. A weak freezing response might reflect difficulty retaining an experience, recognizing a place, or expressing fear through behavior. Each would call for something different to be restored. The tasks the mice struggled with are precisely what can help us think about which jobs a graft would need to take on.

How Much Function Has Dopamine Neuron Transplantation Restored?

A useful comparison comes from dopamine neuron transplantation research aimed at Parkinson's disease. Kriks and colleagues transplanted human embryonic stem cell-derived dopamine cells into animals with damage to the dopamine system on one side. In rats, forelimb stepping and the bias in use of the left and right forelimbs improved. These were improvements demonstrated in specific motor tasks, not evidence of recovery from the disease as a whole or restoration of memory. Kriks et al., Nature 2011 ↗

In a further mouse study by Steinbeck and colleagues, food retrieval that had been biased toward one side became balanced between the two sides after transplantation. But when light was used to temporarily suppress neural activity in the graft, the bias reappeared. A drug that stimulates dopamine receptors could prevent this renewed impairment. The experiment went beyond showing that the graft survived: it linked the improved behavior to graft activity and dopamine signaling. Steinbeck et al., Nature Biotechnology 2015 ↗

Still, it would be inaccurate to divide the approaches into dopamine grafts that “just release a substance” and cortical grafts that “build circuits.” Grafts containing dopamine cells also received input from the host and modulated the function of the remaining striatal circuitry. What matters is which functions have been lost and what remains in the receiving brain.

Here is our reading of that comparison. In models of dopamine system damage, restoring modulation of surviving circuits contributed to improved movement. In the mice in the current study, by contrast, the cells that process information in the neocortex and hippocampus are themselves substantially depleted. Do we expect the graft to help the remaining circuits operate, or to take on part of the work of receiving, retaining, and using information? Without considering that distinction, we cannot judge success simply by the number of transplanted cells or their survival rate.

Considering Both Graft and Host When Asking Why Compensation Was Incomplete

The grafts in this study were still immature and did not have the same layered architecture or cellular composition as mature cortex. Although they connected to the host, it remains unknown whether graft activity was necessary for each behavior. Original paper: discussion ↗

For example, even if some aspects of paw-movement timing can be changed, coordinating the left and right, or front and hind paws, requires inputs and outputs that bring them together. Likewise, adding cortical cells does not necessarily compensate for the sequence of processes, including those involving the hippocampus, needed to associate a place with an unpleasant experience. Rather than reading the results as “human cells lack that ability,” it is more constructive to ask whether the necessary combination of cells and connections was in place.

Allowing the graft to mature further is one possibility. But waiting for maturation may not solve the problem. As the mouse brain develops ahead of it, the conditions under which transplanted cells form connections may also change. Conversely, surviving host circuits might take on the work in another way and influence how the graft is used. These are hypotheses prompted by the results, not established causes.

That makes temporarily silencing graft activity, as in the earlier dopamine transplantation work, an important step. Does performance change when the animal explores a maze or responds in a remembered place? By separating these effects from effects on sensation or movement itself, researchers could test whether graft activity is necessary for that task. Even if it is necessary, further work would be needed to determine whether the graft processes the information itself or modulates activity in host circuits. If performance is preserved when graft activity is silenced, possibilities such as support for host development or reorganization could be examined.

The Connection to Epilepsy: Can Remaining Circuits Be Helped to Function Again?

Transplantation research in epilepsy offers another angle on this question. Hunt and colleagues transplanted embryonic mouse-derived precursors that develop into inhibitory neurons into the hippocampus of adult mice that had already developed epilepsy. They reported fewer seizures and improved spatial learning. In the same study, transplantation into the amygdala did not reduce seizures. The outcomes differed depending on conditions that included the transplant location. Hunt et al., Nature Neuroscience 2013 ↗

This improvement does not mean that lost memories were written back into new cells. Modulating activity in the remaining circuits may instead have helped bring them into a state more suitable for learning. Together with the dopamine transplantation findings, this suggests the possibility that transplanted cells can support function in combination with surviving circuits, without reproducing the lost structures exactly.

Epilepsy, however, is diverse, and not every case can be explained by cell loss. The cortical organoids in the current study also differ in composition and role from the inhibitory cells used in epilepsy research. Suppressing seizures, being able to learn anew, and recovering lost memories must be tested separately. The current findings cannot be directly translated into recovery of a brain affected by epilepsy.

From Surviving Tissue to the Work That Tissue Performs

The original paper also includes an experiment exposing animals to low oxygen after transplantation and examining the response of the human tissue and changes in walking. This was not an experiment that repaired damage after it occurred, but an attempt to study injury to human tissue in relation to the behavior of the animal. Because the exposure affected the whole body, the graft alone cannot be identified as the cause. Even so, the approach offers a useful starting point for disease research. Original paper: low-oxygen exposure experiment ↗

What makes this study interesting is not simply that human tissue grew large. Comparing functions that change after transplantation with those that remain uncompensated lets us ask more specific questions about the cells and circuits required. As dopamine transplantation studies have shown, establishing that graft activity supports a particular behavior can also make it clearer what needs to be improved next.

Beyond that lies another question: can the same graft do the same work in a developing brain and in a brain damaged after reaching maturity? Must the original wiring be reproduced exactly? Or could new combinations with the remaining brain compensate for some functions? Rather than anticipating a treatment, we want to read this as basic research that examines those conditions one by one.