Record-Setting Brain Organoid Holds Vast Medical Potential

By Deborah Borfitz  

October 6, 2026 | Science is at a “magical moment in time” where advances in brain organoid technology and artificial intelligence (AI) are converging to both speed up and derisk the entire drug development process. Human knowledge is poised to take major leaps forward with the discovery of molecular and electrophysiological disease fingerprints in laboratory-grown organoids, models of the human brain. This will change how lab-based drug testing is approached as well as usher in an era of “clinical trials in a vial” to model how the brain, and eventually the entire body, will respond to treatments before people enter studies, according to Paola Arlotta, Ph.D., professor of stem cell and regenerative biology at Harvard University.  

In the latest feat with human brain organoids, Arlotta and her colleagues kept the models alive for a record-breaking six years to enable the study of late developmental stages and postnatal brain maturation (Nature, DOI: 10.1038/s41586-026-10877-x). This follows the 2024 introduction of brain “chimeroids,” single organoids containing the cells of many people (Nature, DOI: OI: 10.1038/s41586-024-07578-8).  

The new study looked with single-cell resolution for transcriptional changes in brain organoids during their development in culture, for comparison with endogenous tissue but from an unprecedented period, she says. “As time was passing in vitro, the cells acquired features that were similar to the cells that spent a similar amount of time in vivo.” The two systems were effectively maturing in parallel, reflecting the fact that the organoids are human and develop and mature in the laboratory following a human tempo.   

Researchers also measured DNA methylation, a reliable and widely-studied marker of a tissue’s biological age that has been used to develop epigenetic clocks. When three of these clocks were fed DNA methylation data from the brain organoids over five years, they detected changes consistent with aging in the endogenous tissue, says Arlotta. Synapses and electrical activity likewise changed over time “much like our brain would change as it develops.”  

A second set of experiments was performed to understand how the passage of time prompted changes in the cells in vitro and in vivo. To accomplish this, Arlotta says, the research team first broke apart a very young and very old organoid to retrieve the single cells that composed them. 

Those cells were then mixed and forced to generate a third, new organoid to restart development of the brain, she continues. Two weeks later, the cells that came from the young organoid were just starting to make differentiated cell types, as was expected at this early stage of development. But those from the old organoid, “sitting right next to the young cells, were doing something very different. They skipped ahead ... [and] went straight to producing cells that normally take several months to make in an organoid.”  

The conclusion was that these organoid systems, despite developing in an artificial environment, have “a cellular memory of the time they spent in culture,” says Arlotta. This time-capsule-like capability manifests in them being able to develop faster to stages of brain development that normally would take months.  

‘Battle of Determination’  

Even five years ago, lab-grown human brain organoids typically lived for no more than a few months, says Arlotta, in speaking about more recent breakthroughs in the field. Longer surviving organoids have expanded the value of the models in studying the progression of human development, which continues for decades.   

Modern brain organoids are much more sophisticated than earlier versions because of massive investment by the field on building models that better mirror aspects of human brain biology, she says, notably features that emerge in the brain over an extended period. The trick has been figuring out how best to support their growth and development and start answering questions about the ways the brain progressively changes and reconfigures its neural connections.  

The secret sauce with the brain organoids Arlotta has been working on is based on know-how about how to culture organoids so that they stay healthy and produce all the expected cells, she says. It is often a “battle of determination” since the cultures need constant nurturing—notably, frequent changes to the media they bathe in to remove waste, provide nutrients, and maintain stability.   

In addition to understanding the cell types residing in these longer-living brain organoids, the research team discovered that the large, excitable neurons in the cerebral cortex would become more fragile and start to die with the passage of time, says Arlotta, although the team demonstrated that they persisted through the years in small numbers. To help them mature and survive longer, they adjusted the media composition of the neurons so that they continued to fire “action potentials,” and therefore being active.   

In the actual brain, scientists know that development and maturation rely on this neuronal activity in response to stimuli from the outside world, such as the sound of a mother’s voice or her touch, she adds. Here, the scientists used an electrical proxy of lived emotional experience by including a modification of commercially available neural media. 

This “permissive" media tells neurons in lab-grown brain tissue to fire if they have the capacity to do so, Arlotta explains. It was inspired by what is known from the study of embryos—that is, neurons spontaneously fire starting long before any external sensory input reaches them.  

Human-Specific Biology 

The new cell culture models are relevant for the study of many brain diseases, including neurodevelopmental disorders (e.g., autism spectrum disorder), neurodegenerative diseases (e.g., Alzheimer’s), psychiatric disorders (e.g., schizophrenia), and mood disorders (e.g., depression), says Arlotta. “We have a limited understanding about how they originate ... [and] progress, and we have very limited ability to predict interventions that would alter the development and manifestation of [those conditions], and this is partly due to the fact that the human brain is unique, and we cannot learn enough from animal models of these diseases.”  

Human-specific brain biology requires working with human samples and thus organoids, says Arlotta, pointing to how DNA sequencing of people with and without autism has revealed genetic states that might be associated with the disease, yet too complex to engineer in an animal model.  

But what these genetic changes do to the developing brain nobody knows, she adds, because the brain has already developed by the time a child sees a doctor. Instead, starting with blood samples taken from patients, scientists are reprogramming blood cells to be stem cells that are then pushed to become an organoid so they can watch a reductionist replica of early brain development and learn what neurodivergent processes may have occurred.   

From that knowledge comes data that can inform the design of interventions and disease progression in individual patients. “Even more interesting,” says Arlotta, “is using the organoids to test how that specific brain organoid replica ... responds to perturbations.” 

Additionally, the organoids could be used to “generate vast amounts of data from the very cells of the patients we’re trying to understand,” she continues. That data could feed AI algorithms to start deciphering how cells respond to various genetic and drug interventions. These “virtual cells” would marry organoids to AI simulation and prediction models and, over the next few years, “open a new way to think about how we develop therapeutics more efficiently.”  

Altering Fingerprints 

After so many years spent understanding organoids, developing better protocols for developing the three-dimensional tissues, characterizing them to confirm that they can accurately mimic the brain, and figuring out what is and isn’t possible, “it’s time to use them to answer some fundamental questions .... [about] how the human brain comes together,” Arlotta says. “Nobody knows how that happens because we don’t have access to [living human brain] tissue.” 

Equally important will be using these organoids to better remedy some of the devastating diseases of society, says Arlotta. “There is not a single drug out there for the core symptoms of autism,” she offers as an example. While autism oftentimes has a genetic origin, how it starts and its progression trajectory are unknown.  

Organoids may be used to reveal the fingerprint of diseases—features such as their gene expression profile or their electrical properties—that are different from the ones in organoids created from the blood of unaffected individuals, she says, which together “may represent the starting and end point of a therapeutic pursuit.” They could be used in the lab to test a variety of drugs and therapies to move the fingerprints from the disease to the control pattern.  

Better yet would be using an organoid containing cells of different people to predict who would respond to certain interventions versus others, says Arlotta, alluding to the chimeroids developed in her lab. The reality is that drugs shown to work on the nervous system in clinical trials don’t work on everybody in the real world.  

“Not only is the human brain different from, say, a mouse brain,” she continues. “My brain is also different from your brain because we are genetically different, and, so far, we cannot predict who would respond and who would not ... [without] a very expensive, long clinical trial.”  

Drug development may soon include clinical trials in a vial using organoids from enrolled participants to define the probability that a person may respond to a new medical intervention, says Arlotta. She is particularly optimistic about the possibilities of integrating human organoid models with advanced AI models trained on data from both people and organoids for predicting effective treatment response to different interventions.  

Ultimately, Arlotta says she imagines these “biological avatars,” made across multiple tissues, will allow predicting how the overall body will respond to therapies prior to clinical trials. The convergence of discovery, knowledge, and technologies from different fields will enable quickly reaching goals that a decade ago were thought to take a century.