Relationship between symptom domains in autism spectrum disorders: a population based twin study.

Dworzynski et al. (2009) looked at twins across the United Kingdom. They used parent and teacher ratings to see how autism traits hang together.

The team ran factor analysis on a big set of behaviors. They wanted to know if the three DSM-IV groups—social, communication, and repetitive—really show up in the data.

Five factors came out, but only partly matched the old triad. Social and communication items mostly landed together, while repetitive behaviors formed their own small factor.

The factors were only modestly linked. Shared genes between repetitive and social-communication traits were small, hinting at partly separate roots.

Seiverling et al. (2012) later reviewed brain scans and saw only weak support for the DSM-5 two-domain model. Their mixed picture backs Katharina’s modest links.

Murphy et al. (2014) studied the same twin registry and found extreme autism traits are highly heritable, yet a clinical diagnosis itself is only 21% heritable. This extends Katharina’s genetic split by showing diagnosis and trait level behave differently.

Lee et al. (2024) used network analysis and again found autistic and irritability symptoms travel in separate clusters. Their newer method confirms Katharina’s core message: symptom groups are loosely coupled.

For you in the clinic, loose links mean a child can score high in repetitive behaviors yet show mild social issues—or the reverse. Do not assume one domain predicts the other. Write separate goals for social-communication and repetitive behavior. Track progress in each area, and pick interventions that match the child’s profile, not the label.