Nobody knows what causes autism, a condition that varies so widely in severity that some people on the spectrum achieve enviable fame and success while others require lifelong assistance due to severe problems with communication, cognition, and behavior. Scientists have found countless clues, but so far they don’t quite add up. The genetics is complicated. The neuroscience is conflicted.
Now, a new study adds an intriguing, unexpected, and sure-to-be controversial finding to the mix: It suggests the brains of children with autism contain small patches where the normally ordered arrangement of neurons in the cerebral cortex is disrupted. “We’ve found locations where there appears to be a failure of normal development,” said Eric Courchesne, a neuroscientist at the University of California, San Diego and an author of the study, which appears today in the New England Journal of Medicine.
“It’s been really difficult to identify a lesion or anything in the brain that’s specific and diagnostic of autism,” said Thomas Insel, director of the National Institute of Mental Health, one of several agencies that funded the project. The new study is notable because it applies sophisticated molecular labeling methods to postmortem tissue from people with autism who died as children, which is incredibly hard to come by, Insel says.
“If it’s real, if it’s replicated and it’s a consistent finding, it’s more evidence that autism starts prenatally and only manifests itself when kids start to have trouble with language or social behavior around age two or three,” Insel said. “These kinds of changes in cellular architecture would happen during brain development, probably around the first part of the second trimester.”
The cortex is the thin sheet of tissue on the surface of the brain. We humans have so much of it that it’s folded up to fit inside our skulls, giving our brains their wrinkly appearance. The cortex plays an important role in everything from basic functions like planning movements and making sense of information from our eyes and ears, to more advanced stuff like language and abstract thought.
If you cut a cross-section through the cortex and looked at it under a microscope, you’d see that it has a consistent cellular architecture, with six distinct layers, each inhabited by certain types of neurons with a certain pattern of connections with other neurons. This uniform organization, many neuroscientists think, is what makes the cortex such a powerful and flexible computer.
But that organization appears to be messed up in spots in many children with autism, according to the new study.
Courchesne and colleagues examined post-mortem brain tissue from 22 children who died between the ages of 2 and 15 — half had autism, half did not. The symptoms of those who had it varied from mild to severe. With help from Ed Lein and other scientists at the Allen Brain Institute, the team applied genetic markers that label specific cell types and specific layers of cortex.
In 10 of 11 of the autistic brains, they found patches of cortex that didn’t follow the normal rules. The patches were a few millimeters across (roughly a quarter to half an inch). In some patches, a specific layer was missing. In others, certain cells weren’t there. The details varied from case to case.
The researchers found these abnormalities in the temporal and prefrontal cortex, areas with roles in language and cognition that are — in a very broad and hand-wavey sort of way — relevant to the symptoms of autism. They did not see them in the occipital cortex, a region primarily associated with vision, which isn’t typically disrupted in autism. Nor did they see them in the brains of 10 of the 11 children without autism. (The one child in this group without autism who had patches of scrambled cortex also had a history of severe seizures, which doesn’t exactly explain that finding, but might be relevant, Courchesne says).
“It’s intriguing to find something consistent like this,” said Helen Barbas, a neuroscientist at Boston University who wasn’t involved in the new study. But she’s less sure about what it means.
One popular hypothesis is that autism results from altered connections within or between regions of the cortex. “The cortex is a huge communication system,” Barbas said. “If you have an abnormality in the structure of cortex, it’s going to affect connectivity.” At this point though, it’s not possible to connect the dots between the scrambled bits of cortex described in the new study and the type of altered connectivity Barbas and others have found previously. “It raises a lot of questions, and that’s good.”
Courchesne acknowledges the new study is just a start. The researchers only had access to small chunks of brain tissue, so they can’t say how widespread the disordered patches were in any given person, let alone how common they are overall in the brains of people with autism (or without it, for that matter). For the same reason, it’s not clear yet whether there’s any relationship between the severity — or the type — of autism symptoms and the number or location of scrambled patches of cortex.
What could cause these abnormalities isn’t clear, but Courchesne thinks genetics and environment could both play a role. The trigger could be some relatively common (but currently unknown) thing encountered by pregnant mothers, Courchesne suggests, but different individuals might vary in their genetic susceptibility to it — and in their genetic potential to compensate for it.
The findings might also be consistent with spontaneous gene mutations, which have been implicated by several teams of autism researchers in recent years, says Robert Hevner, a neuropathologist and neuroscientist at the University of Washington. Unlike the inherited gene mutations passed down from parent to offspring, spontaneous mutations occur later, during development.
“As billions of cells in our body and brain are dividing, mistakes get made,” Hevner said. Because those mistakes affect some cells and not others, they can create a mosaic-like pattern of abnormalities. “If there are mutations occurring on a small scale during brain development, we might see some changes like they’re showing here.”
Still, Hevner sees several reasons to be skeptical about the findings. Chief among them is that the researchers haven’t directly shown that the brains of people with autism have cellular abnormalities — they’ve inferred that from their molecular labeling, which targets RNA. That could be problematic in postmortem tissue, Hevner says. “The brain after death is just sitting there stewing in its own juices, and RNA is a highly unstable molecule that’s easily degraded.”
An alternative interpretation for the new findings, Hevner says, is that the patches with missing molecular markers simply correspond to areas where RNA degraded more quickly than in the surrounding tissue. Courchesne and colleagues did experiments to try to rule this out, but Hevner says he’s still not convinced. “I’ve developed a habit of being cautious,” he said.
When it comes to autism research, that’s probably a healthy habit for everyone.