Individuals who carry common genetic variants associated with autism tend to have lower density in the brain’s microscopic wiring, regardless of whether they actually have an autism diagnosis. The research reveals a shared genetic architecture between the likelihood of autism and the microscopic development of the brain. The study was published in the journal Molecular Psychiatry.
Autism is a condition influenced by a vast pool of genetic variations spread across human DNA. Each minor genetic difference has only a tiny effect on its own, but combined, they shape a person’s underlying likelihood of developing the condition. This type of genetic architecture is called polygenic inheritance.
Researchers have documented structural brain differences in autistic individuals for many years. However, much less is known about how the multitude of genes linked to autism might affect the physical structure of the brain in the broader public. Genetic traits often influence physical characteristics across an entire population on a sliding scale.
To answer these questions, scientists look for subtle patterns in large databases of health records. Yuanjun Gu and Varun Warrier, researchers based at the University of Cambridge, led a large team of international scientists to investigate these patterns. They wanted to see if a higher genetic likelihood for autism corresponds with specific measurable differences in brain anatomy.
The researchers analyzed brain imaging and genetic data from two large, independent sources. They examined information from over thirty thousand adults enrolled in the UK Biobank along with data from nearly five thousand children in the Adolescent Brain Cognitive Development study. Because the team used these massive datasets, they were able to look at both fully developed adult brains and still-developing adolescent brains.
The team focused on five specific physical characteristics of the brain. Three of these features described the macrostructure, or the large-scale shape, of the brain. These included the surface area of the outer layer, the average thickness of the cortex, and the mean curvature of the brain’s folds.
The researchers also looked at two microscopic features by tracking how water diffuses through brain tissue during magnetic resonance imaging, or MRI scans. One of these microstructural measures is the “intracellular volume fraction,” which scientists often use as a reliable indicator of neurite density.
A neurite is a projection from the body of a brain cell. These projections include axons, which send electrical signals, and dendrites, which receive those signals. Measuring neurite density gives researchers a sense of how densely packed the communication wires of the brain are in any given area.
To bridge the physical brain scans with genetics, the researchers calculated a polygenic score for every participant. A polygenic score is a single number that summarizes a person’s total genetic likelihood for a particular trait based on millions of different genetic markers. The team then used statistical models to see if higher polygenic scores for autism correlated with different brain shapes or densities.
The results showed a consistent negative association between the polygenic score for autism and overall neurite density. In both the adult and children populations, people who carried more of the common genetic variants associated with autism tended to have less densely packed neurites. The association appeared globally across the brain’s outer layer, known as the cortex.
The researchers also observed this relationship deeper inside the brain. They looked at white matter tracts, which act as the long-distance communication highways connecting different brain regions. High polygenic scores for autism were also associated with lower neurite density across the majority of these white matter tracts.
Researchers also analyzed the brain as a networked system. Certain areas act as highly connected hubs, similar to major transit stations in a railway network. The team found that the genetic association with lower neurite density was more pronounced in these highly connected hub regions compared to less connected outer edges.
The team also tested a major question regarding sex differences. In the general population, boys are diagnosed with autism much more frequently than girls. Some researchers have hypothesized that biological differences in brain structure between the sexes might explain this diagnostic gap.
The team ran their statistical models again, looking for differences between males and females. They did not find statistically significant evidence to suggest that the autism genes affected the brain structures of men and women differently. The results indicate that the sex disparity in autism diagnoses likely does not stem from different genetic effects on basic cortical structures.
While the correlation between genetics and brain structure is robust, the researchers stress that correlation is not causation. To test if the autism genes were directly causing the changes in brain density, the team used an advanced statistical technique called Mendelian randomization.
Mendelian randomization uses genetic data to see if a change in one trait causes a change in another, sort of like a naturally occurring randomized clinical trial. The models provided no evidence for a direct causal relationship in either direction. The researchers suggest that the genetic link likely results from shared underlying biological mechanisms that influence both the development of the brain and behavioral traits simultaneously.
The researchers note a few limitations to their findings. They only analyzed genetic data from individuals of European descent, which was done to reduce the chance of statistical errors caused by population differences. Research will need to expand to diverse global populations to see if these patterns hold up among different ancestries.
Additionally, current polygenic scores only capture a very small fraction of the total possible genetic variance for autism. The specific microscopic associations reported in this study account for only minor shifts in the overall structure of the brain.
Future research will also need to investigate if reduced neurite density is entirely specific to the genetics of autism. Similar genetic links to reduced microscopic brain density have been documented in studies of other conditions like schizophrenia. Determining whether this brain structure trait is unique to autism or a broader signal for general brain development will be the next major step for the field.
The study, “Polygenic scores for autism are associated with reduced neurite density in adults and children from the general population,” was authored by Yuanjun Gu, Eva Maria-Stauffer, Saashi A. Bedford, APEX consortium, iPSYCH-autism consortium, Rafael Romero-Garcia, Jakob Grove, Anders D. Børglum, Hilary Martin, Simon Baron-Cohen, Richard A. I. Bethlehem, and Varun Warrier.