You have three tabs open: "autism signs", "speech delay 2 year old", and "toddler ADHD". Each one fits a little. Each one also fits the others. This page is about why that is expected at age 2, why the label may matter less than it first seems, and what to do with that.
Four labels that look alike in toddlers
At 24 months, these are the names a clinician might reach for:
| Label | What it names | What you might see at 2 |
|---|---|---|
| Autism | Differences in social communication plus repetitive behaviours or intense interests, with sensory differences common | Not pointing to share, less response to name, repetitive play, strong routines |
| Language delay | Speech or understanding behind expectations, with social connection largely intact | Few words, but pointing, shared looks and back-and-forth play are present |
| Global developmental delay | Delay across several areas (language, motor, thinking, self-care) before a specific diagnosis can be made | Late in most milestones, not only speech |
| ADHD | Attention, activity and impulse regulation well outside what is expected for age | Constant motion, hard to engage for more than seconds, impulsive; rarely diagnosed before 4 |
Every one of these can include a child who does not talk, does not settle, and does not seem to listen. The overlap is not a failure of the clinician. It reflects how development often looks at this age.
The label can shift, and that is expected
Two findings matter here.
First, an autism identification made early tends to hold. In a large general-population study, autism identified from about 14 months onward was stable in most children at follow-up, and more stable than any other early diagnostic category. Very few children moved from autism to typical development. 1
Second, the rest of the picture keeps moving. A study that followed 263 children, most with a family history of autism or ADHD, from infancy to age 6 to 12 found seven distinct outcome profiles. Many children developed typically; others showed combinations of autism, ADHD and anxiety traits; and profiles at age 3 only partly predicted profiles in mid-childhood. 2 Genetics tells the same story from another angle: children diagnosed earlier and later load on partly different genetic factors and follow different trajectories. 3
So the practical picture is this: a first label at 2 is a reasonable description of now, not a forecast. Autism and ADHD can occur together, and clinicians may consider both. A "global delay" at 2 is often a placeholder until a child is old enough to assess precisely. None of this means the first assessment was wrong.
Why the biology overlaps
Autism and ADHD do not appear to be two unrelated conditions that happen to co-occur. Studies of brain networks, of rare genetic variants, and of infants at elevated likelihood report some shared neurodevelopmental biology, alongside features that are specific to each. 4, 5, 6 One infant study found that family history of autism and family history of ADHD related to different early sensory patterns, and that early under-responsiveness to sensory input predicted later traits of both. 4
You do not need the details. The point is that "which one is it?" is partly the wrong question at 2, because the answer for many children is "a mix, in proportions that will become clearer".
Why the supports overlap too
This is the part that can lower the stakes. Ask what a clinician might recommend for a 2-year-old with any of the four labels above, and the lists often look quite similar:
- Communication and play support. The 2023 Project AIM review found evidence for social-communication benefits from naturalistic developmental behavioural interventions. Ask a clinician which strategies fit your child. 10
- Sensory and regulation support: predictable routines, movement, fewer demands during transitions.
- Daycare (CPE) coordination, so the adults who see your child all day work from the same plan.
- Support for you. Parent stress is a legitimate target in its own right; a randomized trial of a short parent group program reduced parenting stress with gains that lasted months. 7
In Quebec, a common starting point for these supports is your CLSC and the Agir tôt program, which takes children under 7 on a voluntary basis, free, and without a diagnosis. 9 The assessment queue sorts out the label; the supports do not have to wait for it. See Waiting is not wait-and-see.
What the label does change
Honesty requires saying where the name does matter.
- ADHD treatment questions. Questions about ADHD treatment depend on age and an individual assessment. The umbrella review linked here summarizes benefits and harms across ADHD interventions; discuss the evidence relevant to your child's age with their clinician. 8
- How you explain your child to family and to yourself. A name can help. It does not have to be the final name.
The professions who sort this out, and who does what, are on Who does what. What the report means when it arrives is on Decoding the assessment report.
This week
0 of 3 · Saved only in this browser. Choose one small step.
Sources
- Evaluation of the diagnostic stability of the early autism spectrum disorder phenotype in the general population starting at 12 months (2019) ↩1
- Mid-childhood developmental and behavioural outcomes in infants with a family history of autism and/or ADHD (2026) ↩1
- Polygenic and developmental profiles of autism differ by age at diagnosis (2025) ↩1
- Cortical markers of excitation/inhibition balance and sensory responsivity from infancy in autism and ADHD family history (2025) ↩1 ↩2
- Connectome-based symptom mapping in children with autism and/or ADHD (2025) ↩1
- Rare coding variants confer high ADHD risk and implicate neuronal biology (2025) ↩1
- Navigator ACT for parents of children with neurodevelopmental disabilities: a randomized trial (2026) ↩1
- Benefits and harms of ADHD interventions: umbrella review and platform for shared decision making (2025) ↩1
- Agir tôt program (2026) ↩1
- Autism intervention meta-analysis of early childhood studies (Project AIM): updated systematic review and secondary analysis (2023) ↩1