How Are Brain Organoids Grown? Stem Cell to Network

Brain organoids are grown from human pluripotent stem cells, usually induced pluripotent stem cells reprogrammed from skin or blood. The cells are gathered into small 3D clusters, steered toward a neural fate with signaling molecules, and kept in supportive culture for months, during which they self-organize into layered neural tissue that develops spontaneous electrical activity.

Key takeaways

Most brain organoids start from induced pluripotent stem cells (iPSCs), a technology introduced by Takahashi and Yamanaka in 2006 and extended to human cells in 2007.
Lancaster and colleagues’ 2013 cerebral organoids showed that human stem cells could self-organize into tissue resembling several regions of the developing brain.
Unguided protocols produce diverse, variable tissue; guided protocols use timed signaling cues to produce specific regions, such as cortex, more consistently.
Growth takes weeks for basic structure and months for maturation; organoids develop coordinated network activity over time, as Trujillo and colleagues documented in 2019.
Key limits are variability, cell stress, the lack of blood vessels (which causes a necrotic core) and immaturity relative to the adult brain.
Assembloids, vascularization strategies and slice cultures are among the main approaches researchers use to build more complex, longer-lived tissue.

What is a brain organoid?

A brain organoid is a three-dimensional, self-organizing tissue grown from stem cells that reproduces some of the cell types, structure and activity of the developing human brain. Typical organoids are up to a few millimeters across. They are sometimes called “mini-brains” in the press, but that phrase is misleading: an organoid is a simplified model of early brain development, not a small version of a thinking brain.

The broader organoid field took off in the late 2000s and early 2010s. Sato and colleagues reported in Nature (2009) that single intestinal stem cells could build crypt and villus structures in a dish. In the brain, Eiraku and colleagues (2008, Cell Stem Cell) showed that embryonic stem cells could self-organize into polarized cortical tissue. Lancaster and Knoblich reviewed the rapid growth of these methods in Science in 2014.

Why grow brain tissue in three dimensions instead of a flat dish? In flat (2D) culture, neurons spread out in a single layer on plastic or glass. In 3D, cells can arrange themselves into layers, form ventricle-like cavities, and connect in all directions, much as they do in the developing brain. That extra structure is why organoids are used to study development and disease, and why they interest researchers working on organoid intelligence. For the computing side of the story, see our complete guide to organoid intelligence.

Where do the cells for brain organoids come from?

Pluripotent stem cells

Organoids start from pluripotent stem cells: cells that can become almost any cell type in the body. There are two main sources:

Human embryonic stem cells (hESCs), derived from early embryos. They were used in many early organoid studies.
Induced pluripotent stem cells (iPSCs), made by reprogramming adult cells, such as skin fibroblasts or blood cells, back into a pluripotent state.

The iPSC breakthrough

In 2006, Kazutoshi Takahashi and Shinya Yamanaka reported in Cell that mouse fibroblasts could be reprogrammed into pluripotent stem cells by introducing four transcription factors: Oct3/4, Sox2, Klf4 and c-Myc. A year later, Takahashi and colleagues reported the same feat with adult human fibroblasts, again in Cell. Yamanaka shared the 2012 Nobel Prize in Physiology or Medicine with John Gurdon for the discovery that mature cells can be reprogrammed to become pluripotent.

iPSCs transformed the field for three reasons. They avoid the need for embryos. They allow researchers to grow neural tissue from specific people, including patients with genetic conditions. And they can be expanded almost indefinitely, so one donation can supply many experiments.

Why donors matter

Every iPSC line carries the genome of the person it came from. That makes informed consent, privacy and the scope of permitted use important, particularly when tissue might be used for purposes donors may not have anticipated, such as biocomputing. We discuss these questions in our article on the ethics of organoid intelligence.

How are brain organoids grown, step by step?

At a conceptual level, growing a brain organoid follows the logic of embryonic development. Researchers do not build the tissue; they create conditions in which cells build it themselves. Protocols vary a lot between labs, but most share a common sequence.

1. Expanding the stem cells

Pluripotent stem cells are first grown and checked to make sure they are healthy, genetically stable and genuinely pluripotent. Problems at this stage, such as accumulated genetic changes in the cell line, carry through to every organoid made from it.

2. Forming 3D aggregates

The stem cells are separated and allowed to clump together into small, floating balls of cells, often called embryoid bodies or spheroids. Starting each organoid from a controlled number of cells helps make them more uniform.

3. Neural induction

Next, the aggregates are pushed toward becoming neural tissue. In the embryo, this happens when signals that would steer cells toward other fates are blocked. In the lab, the same idea is applied by changing the culture medium. A widely used concept is “dual SMAD inhibition”, described by Chambers and colleagues in Nature Biotechnology (2009), which blocks two signaling pathways and efficiently converts human stem cells into neural cells. Cells form neuroepithelium, the sheet of neural precursor cells from which the brain develops.

4. Supporting 3D growth

In many protocols, the developing tissue is embedded in a gel-like extracellular matrix that gives it a scaffold to grow against, encouraging the formation of buds of neuroepithelium. Other protocols skip the matrix and keep aggregates free-floating.

5. Agitated culture and expansion

As the tissue grows, nutrient and oxygen supply become limiting. Lancaster and colleagues moved their organoids into a spinning bioreactor to keep the medium circulating. Qian and colleagues (2016, Cell) developed miniaturized spinning bioreactors that allowed many region-specific organoids to be grown in parallel at lower cost. Orbital shakers serve a similar purpose.

6. Maturation

Over the following weeks and months, neural progenitor cells divide and give rise to neurons, which migrate outward to form layers. Later, glial cells such as astrocytes appear. Neurons extend axons and dendrites, form synapses and start to fire.

A typical timeline

Timing differs substantially between protocols and cell lines. The table below gives a rough, conceptual picture rather than a recipe.

Stage	Approximate window	What is happening	How it is checked
Stem cell expansion	Before day 0	Pluripotent cells grow and are quality checked	Pluripotency markers, genetic stability
Aggregation	First days	Cells form 3D embryoid bodies or spheroids	Size and shape
Neural induction	First one to two weeks	Cells adopt neural identity, form neuroepithelium	Neural markers, morphology
Expansion and patterning	Weeks two to six, roughly	Progenitors proliferate, ventricle-like zones and regions form	Imaging, marker staining
Neurogenesis and layering	First few months	Neurons are born and migrate, layers emerge	Cell-type markers, gene expression
Gliogenesis and functional maturation	Several months onward	Astrocytes appear, synapses form, network activity grows	Electrophysiology recordings

What is the difference between guided and unguided brain organoid protocols?

The biggest design choice in organoid growth is how much to steer the cells.

Unguided (self-patterned) organoids

The cerebral organoids described by Lancaster and colleagues in Nature in 2013 are the classic example of a largely unguided approach. After neural induction, the tissue was allowed to develop with minimal external patterning cues. The result was organoids containing regions resembling several parts of the developing brain, including forebrain and retina-like tissue.

The same paper showed why organoids matter for medicine. Using iPSCs from a patient with severe microcephaly (a condition in which the brain is abnormally small), the authors grew organoids that were smaller and showed premature neural differentiation, offering a human model of the disorder that animal models had struggled to reproduce.

Unguided organoids capture rich cellular diversity and interactions between regions. Their main drawback is variability: which regions appear, and in what proportion, can differ from one organoid to the next.

Guided (directed) organoids

Guided protocols add specific signaling molecules at particular times to push cells toward a chosen brain region. Kadoshima and colleagues, working in Yoshiki Sasai’s lab, reported in PNAS (2013) that human embryonic stem cell-derived cortical tissue could self-organize with layered, polarized structure under directed conditions. Paşca and colleagues reported in Nature Methods (2015) a method for producing cortical spheroids that contained functional cortical neurons and astrocytes, grown free-floating without an embedding matrix.

Guided protocols have been developed for cortex, ventral forebrain, midbrain, hypothalamus, thalamus and other regions. They trade diversity for consistency.

Feature	Unguided (e.g. cerebral organoids)	Guided (e.g. cortical organoids or spheroids)
Patterning	Mostly intrinsic self-organization	Timed external signaling cues
Regions produced	Multiple brain regions in one organoid	Predominantly one target region
Cellular diversity	High	Lower, more focused
Reproducibility	Lower, more organoid-to-organoid variation	Higher
Typical uses	Studying interactions between regions, broad developmental questions	Disease modeling, drug screening, standardized experiments, assembloids
Example	Lancaster et al., Nature (2013)	Paşca et al., Nature Methods (2015); Kadoshima et al., PNAS (2013)

For computing applications, where you need many comparable units, guided and highly standardized protocols are generally more attractive. Paşca reviewed the landscape of these approaches in a 2018 Nature review, “The rise of three-dimensional human brain cultures”.

What role do media, matrices and growth factors play?

Without going into recipes, it helps to understand the main ingredients conceptually.

Culture medium is the liquid that surrounds the tissue. It supplies nutrients, salts and buffering, and its composition changes at different stages to match what the cells need.
Signaling molecules (small molecules and proteins such as growth factors) mimic the cues that pattern the embryo. Blocking certain pathways promotes neural identity; adding others promotes particular regions or later cell types.
Growth and survival factors support the proliferation of progenitors early on and, later, the survival and maturation of neurons.
Extracellular matrix gels provide physical support and biochemical signals that help neuroepithelium organize.
Physical environment, including temperature, oxygen and carbon dioxide levels, and the movement of medium, is as important as chemistry.

The key idea is that organoid growth is staged. Each phase uses a different environment, designed to recreate, in simplified form, what cells would experience at that point in development.

How do researchers check quality and reproducibility?

Because organoids self-organize, no two are identical. Rigorous research depends on measuring that variation and excluding failures.

Common quality control checks

Size and morphology. Organoids are imaged regularly. Those that are too small, too large, or lack clear neuroepithelial structure may be excluded.
Marker staining. Antibodies against proteins characteristic of particular cell types (progenitors, neurons of different layers, astrocytes) show which cells are present and where.
Single-cell sequencing. Measuring gene expression in thousands of individual cells reveals which cell types an organoid contains, in what proportions, and how closely they match reference data from human brain tissue.
Functional recordings. Recording with a microelectrode array shows whether neurons fire and whether activity is coordinated. Our explainer on microelectrode arrays covers how this works.
Cell line checks. Confirming the identity and genetic integrity of the starting stem cells rules out a common hidden source of variability.

How reproducible are brain organoids?

Variability has long been the main criticism of organoids. Encouragingly, Velasco and colleagues reported in Nature (2019) that, with an optimized protocol, individual dorsal forebrain organoids reproducibly generated the diversity of cortical cell types, with organoid-to-organoid variation comparable to the variation seen between individual human brains.

A different study added an important caveat. Bhaduri and colleagues reported in Nature (2020) that cortical organoids showed elevated activation of cellular stress pathways compared with primary human cortical tissue, and that this stress impaired the specification of neuronal subtypes. When organoid cells were transplanted into the mouse cortex, stress markers were reduced. The result suggests that standard culture conditions themselves shape organoid biology, and that matching cell types by markers alone is not the same as matching real brain tissue.

Why do brain organoids develop a necrotic core?

Living tissue needs a constant supply of oxygen and nutrients and a way to remove waste. In the body, blood vessels do this. Organoids have no vessels, so molecules must diffuse in from the surface. Diffusion is effective only over short distances. As an organoid grows beyond a certain size, cells at its center receive too little oxygen and nutrition and die, creating a necrotic core.

This limits how big organoids can get and how long they stay healthy, and it can trigger stress responses in surviving cells nearby. It is one of the main reasons organoids remain small.

How researchers work around it

Agitation. Spinning bioreactors and orbital shakers keep fresh medium flowing past the tissue surface.
Slicing. Giandomenico and colleagues reported in Nature Neuroscience (2019) that culturing slices of cerebral organoids at an air-liquid interface improved survival and allowed the formation of diverse nerve tracts, some of which could connect to mouse spinal cord tissue and drive muscle contractions.
Engineered vessels. Cakir and colleagues reported in Nature Methods (2019) that inducing a subset of cells to express a vascular-promoting gene produced vascular-like networks within organoids and improved some functional features.
Transplantation. Mansour and colleagues reported in Nature Biotechnology (2018) that human brain organoids transplanted into the mouse brain became vascularized by host blood vessels and showed improved survival and integration.
Microfluidics. Microfluidic devices that perfuse fluid through or around tissue are another active line of work.

See our glossary entry on organoid vascularization for more on these strategies.

How do brain organoids mature and start firing?

Electrical activity is what makes organoids relevant to computing. Neurons in organoids generate action potentials and form synapses, and over time individual activity becomes coordinated network activity.

From sparse spikes to network oscillations

Trujillo and colleagues, in Alysson Muotri’s lab, reported in Cell Stem Cell (2019) one of the most detailed longitudinal studies of organoid activity. They grew cortical organoids from iPSCs and recorded them on multi-electrode arrays for about ten months. Activity progressed from sparse, isolated firing to periodic, synchronized network bursts and increasingly complex oscillatory patterns. The authors then trained a machine learning model on electrical features from the EEG of preterm human infants and found that it could estimate the developmental age of the organoids from their activity, suggesting that organoid network development follows some aspects of the early human brain’s trajectory.

Quadrato and colleagues reported in Nature (2017) that long-term brain organoids could develop a wide range of cell types, including photosensitive retina-like cells, and that neuronal activity in some organoids responded to light. Sharf and colleagues reported in Nature Communications (2022) high-density electrode recordings from organoid slices that resolved single-neuron activity within oscillating networks, showing that organoids can support functional circuitry that can be studied at fine resolution with high-density microelectrode arrays.

Why maturation matters for computing

For organoid intelligence, maturation determines what a network can do. Immature networks fire in large synchronized bursts, which can drown out finer, information-carrying patterns. More mature networks show a richer mix of activity and more established synaptic plasticity, the property that allows learning. Methods for training networks are discussed in how neurons learn in a dish.

How mature can organoids get?

Even after many months, organoids resemble fetal or early postnatal brain tissue more than the adult brain in many respects. They lack the sensory inputs, body signals and blood supply that shape real brain development. Extending maturation, without the tissue degrading, is one of the field’s central goals.

What are assembloids?

An assembloid is a structure made by combining separately grown organoids, or organoids with other cell types, so that they integrate. The idea is to build complexity by assembly rather than hoping a single organoid will produce everything.

In a widely cited example, Birey and colleagues in Sergiu Paşca’s lab reported in Nature (2017) that fusing spheroids resembling the dorsal and ventral forebrain allowed inhibitory interneurons from the ventral part to migrate into the dorsal part and integrate into its circuits, recreating a key step of human brain development in a dish.

Assembloids combine the reproducibility of guided protocols with the region-to-region interactions of unguided ones. Researchers have used the approach to connect cortical tissue with other brain regions, spinal cord and muscle. For biocomputing, assembloids offer a route to modular, multi-region networks with more defined input and output pathways.

How do grown organoids connect to computing?

Once an organoid shows stable, coordinated activity, it can be placed on an electrode array to be stimulated and recorded. That is the starting point for biocomputing experiments, including reservoir computing approaches like the one described in our article on reservoir computing with brain organoids.

The quality of the organoid sets the ceiling for every computing experiment. Variability between organoids becomes variability between “devices”. Short lifespans become maintenance problems. Immature networks limit what can be learned. That is why Smirnova and colleagues, in the 2023 Frontiers in Science paper that defined organoid intelligence, placed improved organoid production, including larger, more mature and potentially vascularized organoids, at the core of their proposed research program. Our article on the future of biocomputing examines how these biological limits shape the field’s roadmap, and biocomputing vs silicon explains why the energy case depends on solving them.

Summary: how brain organoids are grown

Brain organoids are grown by taking human pluripotent stem cells, often iPSCs, forming them into 3D aggregates, inducing a neural fate, and maintaining them through staged culture conditions while they self-organize. Unguided protocols yield diverse but variable tissue; guided protocols yield more consistent, region-specific tissue. Quality control combines imaging, molecular profiling and electrical recordings. The main limitations are variability, culture stress, the lack of blood vessels and incomplete maturation, which researchers are tackling with bioreactors, slice cultures, vascularization strategies and assembloids.

Understanding how organoids are grown makes clear why experiments with them need careful documentation: the cell line, protocol, age and quality checks of each organoid all shape the results. Recording that provenance alongside every experiment is part of what we are building at Organoid Programming, as described on our platform page.

Frequently asked questions

How are brain organoids grown?

Brain organoids are grown from human pluripotent stem cells, usually induced pluripotent stem cells made from skin or blood cells. The stem cells are gathered into small 3D aggregates, nudged toward a neural fate with signaling molecules, often supported by a gel-like matrix, and kept in moving culture so nutrients reach them. Over weeks to months they self-organize into layered neural tissue that becomes electrically active.

How long does it take to grow a brain organoid?

Recognizable neural tissue forms within the first few weeks, but functional maturation takes much longer. Organoids typically need months in culture before they show complex, coordinated electrical activity. Trujillo and colleagues tracked cortical organoids for about ten months and saw network oscillations become progressively more complex over that period. Exact timing depends on the protocol and the cell line.

What is the difference between guided and unguided brain organoids?

Unguided organoids, such as Lancaster’s 2013 cerebral organoids, rely mostly on the stem cells’ own tendency to self-organize, producing tissue with many brain regions but high variability. Guided, or directed, protocols add specific signaling molecules at defined times to push cells toward one region, such as the cortex, giving more consistent but less diverse tissue.

Where do the cells for brain organoids come from?

Most brain organoids are grown from induced pluripotent stem cells (iPSCs), adult cells reprogrammed into a stem-cell-like state using a method first shown by Takahashi and Yamanaka in 2006. Some studies use human embryonic stem cells. Because iPSCs come from donors, informed consent, privacy and the donor’s genetic background are important parts of organoid research.

Why do brain organoids develop a necrotic core?

Brain organoids have no blood vessels, so oxygen and nutrients reach the inside only by diffusion from the surface. As organoids grow larger, cells at the center can become starved and die, forming a necrotic core. Researchers address this with spinning or shaking cultures, slicing organoids, engineering vascular-like networks, or transplanting organoids into animal hosts.

Are brain organoids the same as a real brain?

No. Brain organoids are millimeter-scale models of early human brain development. They contain many relevant cell types and show spontaneous network activity, but they lack blood vessels, sensory input, a body, and the full architecture and size of a brain. They resemble fetal rather than adult tissue in many respects, and they vary from one organoid to the next.

What is an assembloid?

An assembloid is a structure made by combining two or more organoids, or an organoid with other cell types, so they integrate and interact. For example, Birey and colleagues reported in 2017 that fusing organoids resembling different parts of the forebrain allowed inhibitory interneurons to migrate from one into the other and form connections, modeling a key step in brain development.

How do scientists know an organoid is working?

Researchers check organoids with several quality control methods: tracking size and shape, staining for markers of specific cell types, sequencing gene expression in single cells to compare against human brain reference data, and recording electrical activity with microelectrode arrays. Organoids that are malformed, show the wrong identity, or are electrically silent can be excluded from experiments.

How brain organoids are grown: from stem cell to firing network

What is a brain organoid?