Mapping genetic influences on the infant brain: A chat with Rebecca Knickmeyer

Although most autism research focuses on children, it tends to exclude the very youngest. That’s particularly true in imaging studies, for which participants must lie still in a scanner for long periods of time. Enter the international research collaboration Organization for Imaging Genomics in Infancy (ORIGINs). In 2017, the organization began to collect imaging, genetics and behavioral data from 6,809 children aged 6 years and under — including data from about 4,000 of them during their first year of life  — from 15 sites in five countries.

Weaving those disparate data threads together will enable researchers to ask fundamental questions about how the brain develops and how genetic variants associated with brain conditions such as autism affect that process, says ORIGINs founder and director Rebecca Knickmeyer, associate professor of pediatrics at Michigan State University in East Lansing. ORIGINs, an arm of the Enhancing Neuro Imaging Genetics through Meta Analysis (ENIGMA) project based at the University of Southern California, is not the first group to take this approach, but it is likely the largest: Imaging genetics studies in infants and young children average just 365 participants, according to a review by Knickmeyer and her colleagues, published in Biological Psychiatry in January.

Spectrum spoke to Knickmeyer about the challenges of imaging babies and young children, and the promise and limitations of identifying neurodevelopmental conditions before age 2.

This interview has been lightly edited for length and clarity.

Spectrum: Why has imaging genetics research focused on adolescents and adults?

Rebecca Knickmeyer: Scanning young children is super challenging. Small babies are scanned asleep, which is very easy to do, but as they start getting to 1 or 2 years old, it’s a little harder to get them to nap. Once they get to 5 or 6, they’re being scanned awake, so they have to feel comfortable in the scanner environment.

Also, the newborn brain is a lot smaller than the adult brain. It’s shaped differently. The contrast between gray matter and white matter tissues is different. So there are a lot of technical challenges in terms of how you analyze data from infants and young children.

S: What has ORIGINs been up to since its launch?

RK: For the first couple of years, we had a lot of people who didn’t have existing funding to do genetic analyses on their cohorts. It’s hard to get funding to do genotyping on an individual cohort with 200 or 500 children, because it’s not big enough. In 2020, we received a five-year grant from the National Institutes of Health that is giving us the opportunity to genotype these large cohorts with imaging from infants and young children and harmonize them.

One of the things that’s special about what we’re doing is, instead of having each site run the image processing on their own, it’s all going to a central site. It’s all being processed on the same systems with the same pipeline, and that will hopefully make the final data more comparable.

Until recently, there weren’t a lot of imaging cohorts out there focused on infancy. But there’s been a real explosion over the past decade. So it’s really the right time to do it.

S: Why has it exploded?

RK: There have been advances in the techniques used for looking at these scans. There are now atlases built specifically for the brains of infants and young children. There are groups such as FIT’NG [Fetal, Infant, & Toddler Neuroimaging Group], which is very much focused on addressing technical issues in scanning infants and young children. There’s just a lot more discussion about how to do it well.

There have also been a lot of advances in the image-acquisition technology so that you get good data in a quicker period of time.

S: How do you plan to integrate the genetic and imaging data?

RK: Once we have this harmonized dataset of 6,000-plus children, we plan to do a genome-wide association study (GWAS) to ask: What genes and variants are contributing to these different neuroimaging phenotypes? We don’t know what the heritability of change in these phenotypes is at this age. We don’t know about genetic correlations between different kinds of outcomes.

It would be really interesting to be able to say, “These five distinct structures are all being driven by the same underlying set of genes.” We don’t know answers to really fundamental questions like that.

That’s one side of it. The other side is we can take genes that are robustly associated with autism, or other conditions such as attention-deficit/hyperactivity disorder or schizophrenia, and ask: What does it look like to have those variants early on in life, before any kind of behavioral differences have emerged? The first way we’re going to do that is just look at the genes that we identify through a GWAS and see how they overlap with the genes that have already been identified for autism and other conditions.

S: Isn’t 6,000 people pretty small for a GWAS?

RK: It certainly is. But you’ve got to start somewhere. And it’s much, much bigger than anything that’s ever been done in infants and young children. We’ll be powered to pick up fairly large effect sizes. We’re not going to be able to pick up individual genes with small effect sizes. But we will be able to ask questions about overall heritability and the correlations among different regions, and the overlap with autism, schizophrenia, ADHD and major depression.

S: How do you hope the information coming out of the study will be used?

RK: When we’re talking long term, one of the things that would be really useful is if we know how genetic factors associated with these conditions manifest very early in life, it could lead to early identification that would allow for earlier interventions, earlier accessing of resources.

Right now, we can effectively diagnose autism starting around 18 months at the earliest. But a lot of people are not identified until much later in childhood. And I don’t think genes on their own will necessarily turn out to be enough when it comes to predicting outcomes for an individual. If you’re going to advocate for getting resources or adjusting the environment in a way that improves the child’s flourishing, the combination of genetics with imaging and perhaps with other clinical predictors could really move us forward.