The Lilliputian versions of human brains that scientists have grown in lab dishes have developed distinct structures such as the hippocampus, grown glia and other cells like those in actual brains, and produced a diverse menagerie of neurons that connect with each other and carry electrical signals. Now scientists have grown hundreds of cerebral organoids with the most complex, human-like activity yet: Though only one-fifth of an inch across, or about the size of a pea, the organoids have developed functional neural networks that generate brain waves resembling those of newborns.

The research, reported on Thursday in Cell Stem Cell, represents a significant advance in creating cerebral organoids that mimic human brain development and function.

“This paper stands out because it looks at the properties and activities of large networks of neurons,” said neuroscientist Hongjun Song of the University of Pennsylvania, who uses brain organoids in his research on brain development but was not involved in the new study. Earlier research found coordinated activity between pairs of neurons, he said, but this one detected “activity in groups of neurons and the production of oscillations” — brain waves.

advertisement

When a preliminary version of the study was released last year, much of the reaction focused on the similarity between the electrical oscillations in the organoids and those in newborns as shown on EEGs. But now that experts have read the formal paper, they emphasize that these brain organoids are not micro versions of newborns’.

“It’s a very solid and incredibly important piece of research,” said Jeantine Lunshof of the Wyss Institute for Biologically Inspired Engineering, who is co-leading a National Institutes of Health-funded study of the ethical issues raised by cerebral organoids. “But it would be wrong to say that these organoids are [scaled-down] baby brains. There is a lot more going on in real human brains than EEGs measure.”

“Are there limits to how accurate we want to make these things? ”

Insoo Hyun, Case Western Reserve University

Still, while creating brain organoids capable of activity similar to actual brains’ might one day reveal how neurological disorders including autism and schizophrenia arise, and possibly inspire ways to treat or prevent them, it also takes neuroscientists into uncharted ethical waters.

“Are there limits to how accurate we want to make these things?” asked bioethicist Insoo Hyun of Case Western Reserve University, who is co-leading the ethics study with Lunshof. “Do you just throw them away” after a study is done? “These are issues that need to be on people’s radar screens and discussed.”

Despite their small size, the cerebral organoids generated several types of brain waves, according to electrodes placed at multiple spots on them. Their neurons fired at frequencies indistinguishable from those that animate actual brains, including gamma waves, alpha waves, and delta waves.

“If you’d asked me five years ago whether we could get organoids to generate sophisticated brain waves, I would have said no,” said biologist Alysson Muotri of the University of California, San Diego, who led the research. “But what we got is unprecedented. No one has ever seen this level of complexity in cerebral organoids, which is why we were so surprised.”

Muotri is a co-founder and part owner of Tismoo, a Brazil-based personalized medicine company that hopes to create brain organoids from the cells of people with autism and rare neurological disorders, and then test experimental therapies on the organoids.

He attributes the organoids’ ability to produce brain waves to their maturity, which is a result of how he and his colleagues adapted a protocol for creating cerebral organoids that scientists in Europe pioneered six years ago. That protocol, or biological recipe, starts with adult human skin cells that are turned into embryo-like stem cells, or induced pluripotent stem cells. Biochemicals then coax these iPSCs to become brain cells, which self-organize into a three-dimensional structure. Muotri and his team spent two years tinkering with the recipe, in particular the nutrient broth where their hundreds of cerebral organoids grow. “By optimizing the medium conditions, we were able to keep them alive longer, up to several years, which gives the neurons time to really mature into complex networks,” he said.

And mature they did. Progenitor cells in the days-old organoids turned into support cells, called glia, as well as neurons. Those grew dendrites, branch-like structures that receive input from neighboring neurons. Although the neurons all initially ran on one neurochemical (glutamate), after a few months the organoid sprouted neurons that ran on others (GABA), just as in the developing brain. They never produced every kind of cell in actual human brains, however.

It was the organoids’ behavior more than their composition that makes this research stand out. Electrodes on the organoids’ surfaces detected synchronous activity from distant neurons, “which showed that different regions are talking to each other,” Muotri said. This activity — which becomes less random as well as more intense, more complex, and more synchronized as the organoids aged, just as they do in the developing brain — produces the waves that EEGs detect.

At first the waves were a single frequency, as in fetal brains, and infrequent, with a frequency of one every 20 seconds in two-monthold organoids. But by the time the organoids were six months old their neural networks — neurons that fire together — produced a symphony of frequencies with irregular but synchronized and more frequent firing, roughly two or three per second, higher than reported by any other lab studying cerebral organoids. That firing rate indicated that the neurons had formed large numbers of functional synapses; in actual brains, each neuron can form thousands, a level of complexity that appears in older fetuses.

Brain waves at frequencies greater than one per second are a hallmark of the human brain.

The brain waves also alternated between quiescence and network-synchronized firing, which “resembles electrophysiological signatures present in [the EEGs of] preterm human infants,” the scientists wrote.

That similarity wasn’t only in the eyes of the scientists. They trained an artificial intelligence system on 567 EEGs recorded from 39 infants born after 24 to 38 weeks of gestation (39 weeks is full-term). The brain waves in 9-month-old brain organoids was comparable to that of a nearly full-term baby, with younger organoids displaying activity patterns like those in premature newborns. “What we have in a dish is following the trajectory of human brain development,” Muotri said.

The brain organoids were not connected to the outside world; no sights or sounds or other stimuli got in. Their neuronal activity was therefore spontaneous and self-generated, as is also true of the proto-brains in the earliest fetuses.

Even before the UCSD scientists reported on how advanced their cerebral organoids became, the mini-brains had already achieved a degree of fame: Last month a batch were flown to the International Space Station for an experiment investigating the effect of microgravity on brain development.

The UCSD brain organoids don’t seem to become any more complex after about nine months. That could reflect the fact that when organoids grow much beyond pea-size, their centers die because oxygen cannot diffuse that deep into the ball of cells; that limits their volume and, probably, their complexity. But Muotri and his colleagues are developing what he calls “an interesting protocol” for supplying the organoids with nutrients and oxygen, basically by surrounding them with living, functioning blood vessels.

If that works, the organoids might grow larger than peas and develop even more lifelike traits. That’s not part of the current paper, but Muotri said he expects to announce it soon.

In the meantime, he is organizing a meeting for October of biologists and philosophers to discuss one of the profound enigmas raised by brain organoid research: How can scientists tell if the organoids’ neural networks produce consciousness?

“I do think we need to pay attention to the ethical consequences of human neural organoids,” said law professor Henry Greely of Stanford University, co-author of a 2018 paper on the ethics of these brains-in-a-dish. If organoids acquire the ability to perceive things “we’d have to worry about pain or other negative feelings,” he said, but that worry would be more akin to concerns about animal welfare and not “humanness.”

That’s because no lab is “anywhere near making human brains in vitro,” Greely added: today’s cerebral organoids have about 2 million neurons, not humans’ 86 billion, and lack specialized regions such as the midbrain and cerebellum. The organoids therefore “just can’t, as far as I can guess, be human,” he said.

Penn’s Song agreed: today’s cerebral organoids “are not really reaching the threshold where they’re brains in a dish.”

But that’s no longer an impossibility. At the upcoming meeting, Muotri said, “we want to talk about how you’d design experiments to prove or disprove the presence of consciousness.”

Leave a Comment

Please enter your name.
Please enter a comment.

  • You can’t talk seriously about determining whether a brain organoid has consciousness or feels pain. That’s like asking whether it has a soul, albeit with different words. The electronic neural networks used today for pattern recognition and deep learning have training signals to indicate correct and incorrect responses. Do they feel pain when they mistake the letter F for the letter E? I don’t think it makes sense to even consider that question or whether it’s murder to update the software on Siri or Echo.

A roundup of STAT’s top stories of the day in science and medicine

Privacy Policy