The autism criteria in the latest version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) lists sensory traits under the umbrella of restricted and repetitive behaviors (RRBs), but this categorization does not reflect how these traits and behaviors present in autistic children, according to a study published yesterday in the Journal of the American Academy of Child & Adolescent Psychiatry.
Instead, RRBs and sensory traits each comprise multiple dimensions, some of which likely share underlying mechanisms, the investigation of more than 1,300 children found.
“Which behavioral domains are included in the DSM-5 criteria, and how they are organized, shapes almost every aspect of how autism is researched, assessed and clinically supported,” says study investigator Emily Spackman, a research fellow at the Kids Research Institute Australia.
For example, behaviors that are considered a single domain don’t count individually toward an overall autism diagnosis, Spackman says. Meanwhile, in clinical cohorts, the DSM-5 categories drive how participants are grouped and which traits are measured individually versus which are collapsed into the same dimension, which can affect results, she says.
On a deeper level, the categorizations influence what “fine-grained clinical domains” the field focuses on, whether considering categories as treatment outcomes or as targets of neurobiological investigations, says study investigator Mirko Uljarević, senior principal research fellow and professor at the Kids Research Institute Australia. “It’s not purely a semantic discussion,” he says.
There has already been “some recognition” in the field that RRBs and sensory traits are related but separate constructs, says Adam Guastella, professor of child and youth mental health at the University of Sydney and the Children’s Hospital at Westmead, who was not involved in the new work. But the study “provides nice evidence” that they can be separated.
“It’s a really important area of research to constantly look at the reliability and fit of the construct for the diagnosis,” he says.
W
The new study probed the accuracy of this categorization using caregiver-reported data for 1,368 autistic children aged 2 to 17 from the SPARK research cohort. (SPARK is funded by the Simons Foundation, The Transmitter’s parent organization.) The study used the Dimensional Assessment of Restricted and Repetitive Behaviors, which includes 98 items divided into eight subscales, and part of the third edition of the Sensory Experience Questionnaire, including 34 items divided into three subscales—sensory hypersensitivity, sensory hyposensitivity and sensory seeking.
A model based on the DSM-5’s four-part definition of RRBs showed the poorest fit with the data, compared with three other models based on the literature. Only one model, which comprised seven RRB dimensions and three sensory-related dimensions, showed an adequate fit, according to four statistical indices.
The results align with empirical findings, says Teresa Tavassoli, associate professor of psychology at the University of Reading, who was not involved in the new work. Her own research suggests that each sensory trait should be broken down further into two kinds: tactile and auditory.
“You can’t just say, ‘This child has sensory differences,’” she says. “You really need to dive deeper.”
Further refinements to the best-fitting model ultimately linked sensory hypersensitivity to insistence on sameness, restricted interests, unusual interests, repetitive language, compulsions and self-injurious behavior in one group and separately linked sensory seeking and hyposensitivity with repetitive motor behaviors.
This grouping tracks with experimental findings, as well as work that has found that sensory hyperreactivity is linked to anxiety and mediated by intolerance of uncertainty, “which is kind of this insistence on sameness; you need to have predictable events,” Tavassoli says. The clustering also makes common sense: “Spinning and spinning an object could be repetitive motor behavior … or because you find it visually pleasing, and it’s more sensory seeking.”
But the new results should be tested against clinical observation and self-report, she adds.
It’s also important to track these dimensions over time to truly distinguish “whether they overlap in potential causes and treatment,” Guastella says. Still, in terms of refining how the field understands these traits, he says, “as a first step, it’s very interesting.”
K
nowing how subdimensions of RRBs and sensory traits might be related is “extremely helpful for both clinical and even etiological research,” Uljarević says. For example, his team hypothesizes that a combination of anxiety and executive function difficulties might underpin the insistence on sameness group, whereas sensory hyposensitivity, sensory seeking and repetitive motor behaviors might share underlying difficulties in regulating arousal.But the larger takeaway is that these traits are all more accurately thought of, studied and treated as dimensions that appear across different brain conditions and also in people who might not qualify for a diagnosis at all, Uljarević says, an approach that Guastella has also supported.
The U.S. National Institutes of Health has attempted to encourage such a transdiagnostic approach by establishing the Research Domain Criteria as an alternative to the DSM categories. It doesn’t help that few instruments are truly transdiagnostic, and collecting a transdiagnostic cohort “poses a burden of having to have multidisciplinary teams,” Uljarević says.
Despite the barriers, he says, such an approach “is essential and fundamental to understanding the mechanisms behind clinical presentations.”
