H. Isaac Chen and the Human Brain Organoids That Wire Into Injured Visual Cortex
Can a piece of lab-grown human brain tissue fill a hole in an injured adult brain and start to process what the eyes see? A team at the University of Pennsylvania led by neurosurgeon H. Isaac Chen, with Hongjun Song and Guo-li Ming, showed that it can in rats. Human forebrain organoids placed in a large cavity of adult rat visual cortex survived, were vascularized by the host, connected to the retina through the visual thalamus, and contained neurons that responded to light and to the orientation of moving bars. The paper is Dennis Jgamadze and colleagues, Structural and functional integration of human forebrain organoids with the injured adult rat visual system, Cell Stem Cell 30(2), 137-152, February 2023 (doi:10.1016/j.stem.2023.01.004).
The study takes the question of replacing brain tissue from single cells to a whole block of tissue, and from a small selective lesion to a large injury. It is the closest current animal model for restoring cortex after brain damage.
A whole-thickness injury and an organoid graft
The team grew human forebrain organoids from stem cells for 80 to 88 days, long enough for rudimentary cortical layers to appear. In young adult rats, they removed the full thickness of a patch of visual cortex by aspiration and placed an organoid in the cavity at once. The rats received daily cyclosporine to suppress rejection of the human tissue.
- Grafts survived at one, two and three months, with an overall survival rate of 82.1% for the main stem cell line.
- Host blood vessels grew into the organoids, and the number of vascular structures was established by one month and increased after that.
- Whole organoids survived and grew far better than grafts of cells dissociated from organoids.
Connected to the eye
The authors traced connections in both directions with engineered viruses. To test whether visual information reached the graft, they injected a herpes virus that crosses synapses forward into the eye on the side opposite the graft. The virus reached the organoid neurons by way of the lateral geniculate nucleus, the visual relay of the thalamus. The authors describe this as unequivocal evidence of a polysynaptic pathway from the host retina into the graft. Graft neurons also sent axons out to other parts of the host visual system.
Responses to light and orientation
The team recorded single units in the grafts with multi-electrode probes while the rats viewed a flashing screen and moving gratings.
| Measure | Organoid graft | Naive rat visual cortex |
|---|---|---|
| Animals with visually evoked activity | 6 of 8 (0 of 1 at one month, 4 of 5 at two months, 2 of 2 at three months) | 2 of 2 |
| Graft units evoked by light, two months | 31 of 140 (22.1%), plus 10 suppressed | Not applicable |
| Graft units evoked by light, three months | 7 of 33 (21.2%) | Not applicable |
| Units with evoked activity, comparison | Fewer than in naive cortex | 22 of 30 (73.3%) |
| Orientation selectivity index distribution | Not statistically different from naive cortex (p = 0.649) | Reference |
Orientation was tested in four rats, two at two months and two at three months. Some graft units fired preferentially for gratings of one orientation, which is a feature of visual cortex processing that goes beyond a simple response to light.
Why tissue structure helps
The discussion section compares this result with earlier grafting work. Single neurons placed into a small selective lesion, as in the transplanted neuron study by Magdalena Götz and Mark Hübener, reach orientation selectivity when the cortical architecture around them is intact. Grafts of dissociated cells placed into large lesions in earlier studies responded to stimuli but showed no more complex responses. The authors conclude that intact cortical architecture helps graft neurons take on higher-order function, and that larger lesions may need a graft that brings its own structure, which they describe as an auxiliary cortical column.
A related study from Sergiu Pașca’s lab at Stanford placed human cortical organoids into the intact cortex of newborn rats (Revah et al., Nature 2022, doi:10.1038/s41586-022-05277-w). The grafts received thalamic and cortical inputs, responded to sensory input, and their optogenetic activation could drive reward-seeking behavior. That work did not involve an injury, which is the step the Penn study adds.
What this tells us about consciousness
The question of restoring consciousness after brain damage has two parts. One is to drive the circuits that remain. The other is to give the brain new tissue that can take on lost functions. This study is evidence that the second part is possible for sensory processing, using human tissue grown from stem cells that never belonged to the host. The organoid neurons had no visual history and no link to the rat’s past. Their responses came from the inputs they received after they were placed.
That supports the idea of a neutral template on which a person’s own activity writes the function, which our Neutral Core page sets out. It also says something about the limits of the idea. Orientation tuning in visual cortex is shaped by the input from the eyes, so a graft can relearn it as long as the input arrives. A function that depends on stored history, such as a memory, has no such ongoing teacher. The experiment does not test that.
Brain organoids also raise their own welfare questions, which our coverage of Alysson Muotri’s organoid consciousness criteria addresses. A graft that integrates with a living brain is no longer an isolated dish culture, and any assessment of it has to include the host. The broader questions about gradual transfer and substrate are collected on our brain emulation page.
Comparison to The Consciousness AI
The Neutral Core’s slow handover section cites this paper as the route for restoring function in injured cortex. The Core itself is a simulation of leaky integrate-and-fire neurons, of the kind a reader can run in the Substrate Console, and nothing in the project has tested it against living tissue. The study gives two concrete benchmarks that any synthetic replacement would need to meet. It should receive input that traces back to the sense organs, and its units should develop response properties statistically indistinguishable from the tissue they replace.
Limits of the organoid repair result
H. Isaac Chen’s team showed that human brain organoids can integrate structurally and functionally with the injured adult rat visual system. The limits are clear and the authors state most of them. Immunosuppression with cyclosporine limited follow-up to three months. Fewer graft neurons responded to light than neurons in normal cortex. Orientation tests covered four rats. An aspiration cavity is a severe injury but does not reproduce the biology of stroke or traumatic brain injury. No behavioral recovery was tested. Within those limits, the study shows that a block of new human tissue can join a damaged brain and process its sensory input, and it belongs with the measurement questions in how scientists race to define and test AI consciousness in 2026, since a repaired brain has to be shown to work like the original.
Researchers covered here
- H. Isaac ChenDepartment of Neurosurgery, Perelman School of Medicine, University of Pennsylvania, and Corporal Michael J. Crescenz VA Medical CenterTransplanting human brain organoids into injured adult cortex, structural and functional integration of organoid grafts with the visual system
- Sergiu PașcaStanford UniversityHuman cortical organoids and assembloids, the Brain Organogenesis program, xenocortical mice