In one sentence
Researchers piloted small wrist-worn accelerometers during functional analyses with three autistic children who self-hit, comparing sensor data with live observation and frame-by-frame video coding.
What the researchers did
A functional analysis (FA) is a structured assessment used to understand why a challenging behavior happens so clinicians can design function-based treatment. Data collected during FAs depend heavily on observers, which leaves room for human error.
In this pilot study, three autistic children who engaged in self-hitting wore accelerometers — small devices that record movement — while researchers conducted an FA. The team compared three measurement approaches: live observation in the session (“clinical-grade”), frame-by-frame video review (“research-grade”), and accelerometer readings. They then calculated interobserver agreement across the three data sets and discussed practical recommendations.
What they found
This was an early feasibility study, not a large trial. The abstract reports that agreement was calculated across observation, video, and sensor data, with discussion of how the methods compared and what that might mean for practice.
Because only three children participated and the published summary does not report detailed outcome numbers, the findings should be read as preliminary. The study focused specifically on self-hitting rather than other forms of self-injury, and sensors were worn on the wrist.
What this means for families and therapists
For behaviour analysts and ABA teams, wearable sensors may eventually offer a useful second data stream during functional analyses, especially when self-injury is fast or easy to miss in real time. That could support more reliable treatment planning — but only if the technology is validated for the specific behaviour and setting.
For families, this research is not yet a home product recommendation. It is a signal that clinicians are exploring tools that may reduce reliance on human observation alone. If a team proposes sensor-based measurement, it is reasonable to ask how the data will be combined with clinical judgement and whether the method has been tested for your child’s specific behaviour.
Limitations and what we don't know yet
Only three participants were included, so results may not generalise to other autistic children or other forms of self-injury.
The study tested detection during a structured FA, not whether sensors improve long-term treatment outcomes. Sensor placement, movement type, and setting all matter, and false positives or missed events remain possible. More research is needed before this approach becomes standard practice.
This is a simplified summary of Wearable Technology to Measure the Occurrence of Self-Injury During a Functional Analysis by Neely L, Holloway K, Miller S, Cantero K, Alaeddini A, Oyama S, Behavior modification (2025).
Source license: CC-BY-NC.
This is not medical advice. Please consult a qualified healthcare professional before making treatment decisions.