Novel clustering of items from the Autism Diagnostic Interview-Revised to define phenotypes within autism spectrum disorders.
Four data-driven ADI-R phenotype clusters give BCBAs a sharper picture for treatment planning and research matching.
01Research in Context
What this study did
The team fed every answer from the Autism Diagnostic Interview-Revised into a computer.
About 2,000 autistic people’s scores were sorted with cluster math.
The goal was to see if clear-cut behavior groups pop out, ready for gene studies.
What they found
Four natural clusters appeared.
One group had big language delays.
Another showed mild problems across the board.
A third had many savant skills.
The last sat in the middle on most items.
How this fits with other research
Bitsika et al. (2018) saw only high- and low-severity groups, not four flavors.
The gap is mostly age and items: they used school-age kids and different checklists, so both can be right.
Georgiades et al. (2014) kept the same ADI-R items but tracked the same kids until age six.
They found a small slice whose symptoms shrank over time, showing the clusters can move.
Sacco et al. (2012) also landed on four groups, but used medical history traits instead of interview answers.
The four-cluster picture now repeats across labs, tools, and countries.
Why it matters
You can stop saying “autism is just a spectrum.”
Think in four rough lanes when you read an ADI-R report.
If a child lands in the severe-language cluster, prioritize verbal behavior targets.
If the profile shows savant skills, build sessions around those strengths to keep motivation high.
Share the cluster label with medical partners so genetic and intervention studies compare apples to apples.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Open the last ADI-R you filed and tag which of the four clusters best fits your client, then adjust language and reinforcement plans to match that profile.
02At a glance
03Original abstract
Heterogeneity in phenotypic presentation of Autism spectrum disorders has been cited as one explanation for the difficulty in pinpointing specific genes involved in autism. Recent studies have attempted to reduce the "noise" in genetic and other biological data by reducing the phenotypic heterogeneity of the sample population. The current study employs multiple clustering algorithms on 123 item scores from the Autism Diagnostic Interview-Revised (ADI-R) diagnostic instrument of nearly 2,000 autistic individuals to identify subgroups of autistic probands with clinically relevant behavioral phenotypes in order to isolate more homogeneous groups of subjects for gene expression analyses. Our combined cluster analyses suggest optimal division of the autistic probands into four phenotypic clusters based on similarity of symptom severity across the 123 selected item scores. One cluster is characterized by severe language deficits, while another exhibits milder symptoms across the domains. A third group possesses a higher frequency of savant skills while the fourth group exhibited intermediate severity across all domains. Grouping autistic individuals by multivariate cluster analysis of ADI-R scores reveals meaningful phenotypes of subgroups within the autistic spectrum, which we show, in a related (accompanying) study, to be associated with distinct gene expression profiles.
Autism research : official journal of the International Society for Autism Research, 2009 · doi:10.1002/aur.72