Researchers have grown miniature, three-dimensional brain models from the cells of autistic and nonautistic individuals to study how their neural networks communicate. The models demonstrated that different types of autism spectrum disorder produce vastly different patterns of electrical activity, highlighting the wide biological variety underlying the condition. The small study was published in the journal Translational Psychiatry.
Autism spectrum disorder is a neurodevelopmental condition involving differences in social communication and repetitive behaviors. The biological roots of the condition remain difficult to map out. While many cases have no known genetic cause, a portion of autistic individuals have what is known as syndromic autism. This form of the condition is linked to specific single-gene mutations that alter how brain cells develop and communicate.
Because traditional animal models often fail to accurately reflect the specific features of human brain development, researchers have increasingly turned to brain organoids. These are tiny, self-organizing bundles of tissue grown from human stem cells. Brain organoids replicate the early stages of human brain development while retaining the exact genetic code of the person who provided the original cells.
Lead researchers Nisim Perets and Liya Kerem, along with a team of colleagues at Itay and Beyond and the Hebrew University of Jerusalem, wanted to see if organoids could reveal the functional differences between various forms of autism. They focused on comparing the baseline electrical activity and network connections among several distinct genetic subtypes of the disorder. Extrapolating how a single mutated gene alters whole-brain activity is challenging in living humans, making these laboratory-grown models highly useful for observing live neural networks in action.
The researchers collected urine samples from fifteen human participants. Four participants were neurotypical, serving as a control group. Ten participants had syndromic autism stemming from five different genetic mutations, including the genes SHANK3, SCN2A, STXBP1, PPP2R5D, and GRIN2B. One participant had idiopathic autism, meaning their condition had no identified genetic origin.
Using epithelial cells extracted from the urine, the team reprogrammed the cells back into a basic stem cell state. They then placed these induced pluripotent stem cells into special nutrient baths, guiding them to grow into more than four hundred brain organoids. To ensure the models were developing correctly, the researchers analyzed the cells using genetic sequencing and fluorescent imaging. This confirmed that the organoids contained the right mix of brain cells, including neural progenitors, developing neurons, and mature cortical cells. After growing the organoids for about two months, the researchers placed them onto special plates equipped with microscopic electrodes. These multi-electrode arrays allowed the team to record the spontaneous electrical signals passing between the neurons. The researchers tracked metrics like the firing rate, the size of the electrical spikes, and the frequency of synchronized bursts across the neural network.
The resting electrical activity in the autism models differed substantially from the neurotypical control models. The organoids derived from the participant with idiopathic autism exhibited a generally hypoactive profile. They showed lower firing rates, weaker signal strengths, and fewer bursts of activity compared to the control group.
Add PsyPost to your preferred sources Conversely, organoids derived from most of the syndromic autism subtypes showed higher firing rates than the control group. Organoids from participants with SCN2A mutations had varying firing rates but consistently produced weaker electrical signal strengths. This initial observation confirmed that genetic differences lead to physical differences in how brain cells spontaneously fire.
Next, the research team tested how the neural networks responded to new stimuli, a process known as short-term synaptic plasticity. In a living brain, neural networks adapt to incoming information by temporarily adjusting their sensitivity. They might dampen their activity, known as short-term depression, or temporarily boost it, known as short-term potentiation. The researchers delivered brief electrical pulses to the organoids and recorded the changes in activity over the following five minutes.
The high-frequency stimulation mostly caused the networks to dampen their activity. However, organoids from patients with STXBP1, SHANK3, and SCN2A mutations displayed abnormally high levels of short-term depression and reduced potentiation compared to the control group. Organoids with GRIN2B mutations showed the opposite trend, with slightly elevated potentiation and reduced depression. These varied responses suggest that different genetic mutations disrupt the brain’s ability to adapt to incoming signals in entirely different ways.
The researchers also mapped the functional connectivity of the networks before and after the stimulation. In response to the electrical pulses, the neurotypical organoids displayed a stable, predictable decrease in network connectivity. Organoids from the autism groups displayed highly erratic responses.
The neural networks in organoids with a STXBP1 gene mutation collapsed almost immediately after stimulation, failing to recover normally. Models with a PPP2R5D mutation experienced a sharp, sudden drop in connectivity. Models with GRIN2B mutations showed an inconsistent, fluctuating response across the entire observation period. The organoids representing idiopathic autism barely changed at all, showing a rigidity not seen in the control group.
To visualize these vast differences, the researchers mapped eighteen separate electrical characteristics onto a three-dimensional graph using a mathematical technique called principal component analysis. Organoids grown from the same person behaved similarly, and the entire neurotypical control group clustered tightly together. The organoids from the autistic participants scattered widely across the graph.
Even organoids grown from patients sharing the exact same genetic mutation sometimes displayed different patterns of electrical activity. One participant with a GRIN2B mutation had a clinical history of seizures, and their corresponding organoids exhibited abnormal rhythmic bursting. Another participant with the same GRIN2B mutation did not have a history of seizures, and their organoids did not display that specific bursting pattern.
While these three-dimensional models provide a window into early brain development, they do not replicate the full structural architecture of a mature human brain. The study also relied on a small sample size of fifteen patients, with only a single individual representing the idiopathic autism category. The association between specific electrical patterns in the organoids and clinical symptoms like seizures will require testing in larger patient cohorts to verify that the results were not statistically anomalous. The researchers noted that tracking functional electrical differences in brain organoids could help classify different subtypes of autism based on brain circuitry rather than just behavioral observations. Because autism encompasses such a wide array of biological mechanisms, grouping patients by how their neural networks actually function might eventually guide the development of tailored therapeutic interventions.
The study, “Patient-derived brain organoids reveal divergent neuronal activity across subpopulations of autism spectrum disorder,” was authored by Nisim Perets, Liya Kerem, Nir Waiskopf, Noa Horesh, Itay Goldman, Jasmine Avichzer, Doron Bril, William Tobelaim, Milcah Barashi, Liat David, and Ariel Tenenbaum.