Trajectories of Autism Symptoms and Overlapping Patterns in a Chinese Cohort From 18 to 36 Months.
ADOS-2 severity paths overlap between autism and BAP toddlers through age three, so keep assessing before you finalize.
01Research in Context
What this study did
Zhang et al. (2026) watched 18- to 36-month-old Chinese toddlers.
They used the ADOS-2 every few months to track autism severity.
The team ran latent-class growth modeling to see if autistic kids, broader-autism-phenotype kids, and typical kids traveled different paths.
What they found
The curves overlapped.
Many toddlers with autism and many with BAP landed in the same severity bands at the same ages.
Girls and kids who started with stronger social-affect scores were more likely to end up in the autism group, but the lines still crossed.
How this fits with other research
Liu et al. (2026) looked at the same Chinese cohort and found clear cut-offs: CSS 5 for most kids, CSS 6 if global delay is present.
That seems opposite, but it isn’t—Jacqueline set static thresholds for one-time diagnosis, while Feixia showed movement across time.
Dellapiazza et al. (2024) followed 575 French children and also saw fluid trajectories, giving a cross-country replication.
Kantzer et al. (2018) warned that 1 in 5 screen-positive toddlers shift diagnosis within two years, foreshadowing why single scores can mislead.
Why it matters
For you, this means park the label and watch the curve.
Repeat the ADOS-2, add cognitive and language data, and wait until at least 36 months before you lock the file.
Use Jacqueline’s cut-offs as red flags, not stop signs, and keep caregiver coaching running the whole time.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Schedule a follow-up ADOS-2 in three months for any toddler who scored CSS 4-6 today, and add a language sample.
02At a glance
03Original abstract
Several studies have examined trajectories of autism symptoms in autistic children and their siblings. However, less is known about developmental patterns across different neurodevelopmental conditions, as well as the clinical characteristics and early predictors among children with distinct diagnoses but overlapping trajectories. This study investigated trajectories of autism symptoms [ADOS-2 Calibrated Severity Score; Total CSS, social affect (SA) CSS, and restricted and repetitive behaviors (RRB) CSS] and their overlap across symptom domains in a Chinese cohort of 163 children aged 18-36 months, including autism, broader autism phenotype (BAP), developmental delay (DD), and typical development (TD). Latent class growth modeling identified three trajectories for Total CSS and SA CSS, and two for RRB CSS. Cross-domain analyses revealed overlapping CSS trajectories between 51 autistic children (Overlap-Autism) and 36 children with BAP (Overlap-BAP). Clinical comparisons between these two groups revealed distinct temporal profiles in CSS and ADI-R scores. Hierarchical logistic regression analyses indicated that female sex was a protective factor for Overlap-Autism (OR = 0.20, p = 0.009) relative to the Overlap-BAP group, whereas higher SA CSS at 18 months independently predicted Overlap-Autism (OR = 1.61, p = 0.003). These findings highlight the challenges of early diagnostic differentiation among children with autism-related traits and underscore the importance of comprehensive and longitudinal assessment.
Autism research : official journal of the International Society for Autism Research, 2026 · doi:10.1002/aur.70219