For decades, autism diagnosis has relied on a deceptively simple premise: watch a child, ask a parent, and make a judgment. The same basic approach used in 1943, when Leo Kanner first described the children in his clinic, remains the standard today.
James McPartland, PhD, wants to change that.
In a fluorescent-lit laboratory at the Developmental Disabilities Clinic at Yale Child Study Center (YCSC), McPartland and his team are hunting for something that has eluded psychiatry for generations: biological markers that can reveal autism directly, without the uncertainty of observation, without the expense of expert clinicians, without the wait.
“As clinicians reliant on behavior, we can’t measure things that a child doesn’t do in front of us,” McPartland said. “And we have no way of looking into the future to predict what behaviors are going to emerge.”
The limitations are not abstract. Families in rural America can wait years for a diagnostic evaluation, driving hundreds of miles to see one of a shrinking pool of specialists.
The cost can run into thousands of dollars. And even then, the answer rests on a snapshot—what a child happens to show on a particular Tuesday morning.
McPartland, the Harris Professor of Child Psychiatry and Psychology, came to neuroscience as a clinician frustrated by these constraints. He has written seven books and more than 200 scholarly papers, but his core question remains practical: Can measuring the brain directly do better than watching the child?
The early answers suggest yes.
Across multiple large studies, Yale researchers have established what McPartland calls “truths” about autism.
On average, autistic and non-autistic people look at faces differently. Their brains process facial information differently. Those differences are consistent enough to be measured and significant enough to matter.
An eye-tracking tool based on Yale discoveries from two decades ago is now being explored commercially as a diagnostic aid. Electroencephalography, or EEG, already used in hospitals to diagnose seizures and screen newborns for hearing loss, could potentially be deployed for autism screening at a fraction of the cost of a specialist evaluation.
But McPartland is careful not to overpromise. Developing biological tools for a condition defined by behavior is hard. The field has learned to stop searching for a single autism biomarker and start looking at specific brain systems tied to specific functions.
“We’re shifting from a broad view of autism to a more detailed approach,” he said.
That shift is unfolding in real time at Yale, where neurosurgeons now sit alongside child psychiatrists. In an ambitious new project, researchers are recording electrical activity from inside the brains of epilepsy surgery patients to understand how the brain processes language in autism.
Two recent papers from McPartland’s group, done in partnership with the Yale Biomedical Imaging Institute and the Positron Emission Tomography Center, found molecular differences in the brains of autistic people—different kinds of building blocks, different neurotransmitters, and synapses.
“When I came here in 2004, I wouldn’t have imagined collaborating with neurosurgeons,” McPartland said. “Yale is a unique place with a uniquely collaborative environment.”
The stakes go beyond academic curiosity. Autism now affects one in 31 children, according to the CDC. That figure has risen sharply, not because of an epidemic but because diagnostic criteria have broadened and stigma has fallen.
In the 1950s, when clinicians wrongly blamed autism on cold parenting, families hid. Today, public figures speak openly about their experiences. More parents seek evaluations. More children get diagnosed.
That is progress. But it has strained a system still relying on tools from 1943.
Biomarkers could change the math.
A cheap, scalable EEG screening could tell a pediatrician which children need a full evaluation. An eye-tracking measure could give a clinician confidence in a diagnosis that might otherwise take months to confirm.
For families in places without an autism expert, that is not an abstraction. It is the difference between early intervention and watching a child fall further behind.
McPartland emphasizes that biomarkers are not about replacing human judgment. They are about supplementing it. The goal remains the same as it has always been: helping autistic people live functional, happy, effective lives. The path to that goal, however, may soon look very different.
“We are making unprecedented progress,” he said. “This multidisciplinary approach will prove vital for advancing our knowledge and treatment of autism.”
Now, researchers are involving autistic people in the work and exploring what the neurodiversity movement means for the field. For now, in a lab at Yale, the search for biological truth continues—one EEG, one eye movement, one collaboration with a neurosurgeon at a time.