Meet the parents:

The concept of the broad autism phenotype sprang from landmark research by psychiatrists Susan Folstein and Michael Rutter. In 1977, they conducted the first twin study of autism, analyzing 21 sets of British twins, identical and fraternal, in which at least one sibling had the condition. They showed that autism occurs more often in both twins when the twins are identical, providing strong evidence of autism’s genetic roots. But even when only one twin has autism, the unaffected twin sometimes has delayed language and difficulties with spelling and reading. This finding led Folstein and Rutter to conclude that “autism is genetically linked with a broader range of cognitive disorders.”

Folstein pursued this idea in the early 1990s, conducting some of the first studies of the parents of autistic children. Her team noticed, for example, that some parents of autistic children have mild language difficulties, recalls Landa, who worked with Folstein at the time. “Their conversational turns would be excessively long, or sometimes they had a lot of difficulty formulating their ideas.”

In 1991, the team sat down to compare notes with Joseph Piven, who had also been noticing characteristics such as rigid thinking in the parents of autistic children. Landa remembers Piven describing one father who had refused to mail a letter for his wife on his way to work because it meant he would have to alter his route. “What we realized was that these various traits we were noticing resonated with the characteristics of autism, but they were subclinical and not all family members had all these things,” Landa says.

The researchers described what they were seeing as the “lesser variant of autism,” Landa says. In a series of papers over the next five years, they reported a higher level of autism traits among the parents of autistic people than among parents of children with Down syndrome or typical children. Many of these parents also have an elevated prevalence of psychiatric disorders, especially anxiety, as well as problems with language and social skills. Losh and her team have found many of these same traits in women who have a child with fragile X syndrome, which accounts for approximately 5 percent of all autism cases. The women don’t have fragile X syndrome but carry a ‘premutation’ in FMR1, the gene mutated in the condition. “That some of these [traits in the mothers] look very much like those observed in the broad autism phenotype is exciting because it implicates FMR1 in features related to autism spectrum disorder,” Losh says.

Using an eye-tracking device, her team has also found some less obvious quirks in some parents of autistic children — including telltale patterns of attention and language processing that distinguish these parents from both typical people and autistic individuals.

“We think of [the broad autism phenotype] as the distilled expression of autism’s genetic liability.” Molly Losh

“Boat … star … pencil.” Wiesenthal is calling out the names of objects in an array on a computer screen. She is at Losh’s lab on a wintry day in February to take part in some testing. Her eyes, monitored by the device, move from object to object, left to right and top to bottom, as if she were reading a page in a book. She hesitates for a second, then continues, “… chair … fish … key.” She has already done the same exercise with colors, letters and numbers. “I wonder how my father would do on this,” she says when she finishes.

The skill being tested is called ‘rapid automatized naming.’ It seems simple on the surface, but as people read, their eyes typically jump from one word or object to the next in movements called saccades. On the eye tracker’s monitor, these jumps show up as thin red lines that traverse the screen like unspooling threads. To register the word or object, though, the reader’s eyes must stop briefly on a point of fixation, represented by a red dot. The timing of this response requires the brain to sync sensory input, attention and executive function. “It is an indirect window into our cognitive ability,” says Kritika Nayar, a graduate student in Losh’s lab who leads this work.

In typical people, the eyes lead the way, looking one or two objects ahead of the one being named. Typical people tend to fixate on only one point, and they call out names fluidly, one after the other, as the red dot bounces steadily along. Autistic individuals name the objects on the screen less fluidly: Their gaze skitters around each object before settling on a fixation point, and they perseverate more in general, getting stuck on objects and looking back at previous items. On the monitor, the red dot mostly moves with their voice, not ahead of it.

Parents with characteristics of the broad autism phenotype, meanwhile, fall neatly into a cognitive middle ground. They perform more fluidly than autistic people but get stuck more often and fixate on more points than typical people do — something they would never notice in everyday life. (Losh shares results if participants ask; although those results are significant in terms of group differences, they are not necessarily meaningful for an individual participant.) The broad autism phenotype does not manifest uniformly across people who have it. But because this task is automatic and tied to specific brain circuits, it could be useful as a marker for both autism and the broad autism phenotype, Losh says. It could also help to find various genetic contributors to the condition. “This is a potentially more powerful approach to draw gene-behavior associations” than looking only at autism, she says.