News

New research published in the Journal of Autism and Developmental Disorders has found that artificial intelligence (AI) powered wrist and ankle wearables could transform how clinicians, families and schools detect and respond to escalating distress in autistic children and adolescents, potentially offering advance warning of meltdowns, self injury or emotional overload minutes before they occur. This could be a huge improvement in care.

The research (led by Shabinabegam Abdul Majid Sheth, Emilios Lemoniatis, Tayyaba Ejaz, Eilis Kennedy and Gabriella Charalambous) from the Tavistock and Portman (now part of NLFT), University College London and partner institutions, evaluates the effectiveness, feasibility and acceptability of AI and machine learning enabled wearables designed to monitor physiological and behavioural states in autistic young people. These devices, often resembling simple wristbands or ankle trackers, detect small changes in heart rate, skin response, movement and other body signal that may precede visible distress. 

________________________________________

Seeing the signs before they surface

For many autistic children and their families, behavioural escalation can feel unpredictable and without warning. Yet the research found that the body often makes small but meaningful changes before outward behaviour shifts. Across several trials, these wearable devices detected changes in heart rate, skin response and movement that typically preceded aggressive episodes by one to three minutes, with high accuracy indicating reliable early signals across specialist psychiatric settings. This represents a breakthrough proof of concept. 

Tests have shown that AI enabled wearables can spot signs of self injury and repetitive movements with a high level of accuracy, especially in controlled settings. In some early personalised models, accuracy was even higher. This showed how tailored approaches could help clinicians step in earlier. They were also able to identify different emotional and sensory states in most cases when tested in lab environments. These early results show that physiological shifts are often detectable well before behaviours escalate – a discovery that could reshape how support is provided both at home and in clinical care.

________________________________________

Designed for everyday life

A recurring challenge with technology based monitoring is that many devices that work well in research laboratories simply can’t be worn comfortably or discreetly in real life. This review took a different approach, looking exclusively at tools that resemble familiar items such as watches or soft bands. 

As psychiatrist Dr Shabina Sheth noted, the true test of these devices is whether a child can wear them ‘in school, at home, or during sensory overwhelm’, not just in a controlled environment. 

Across the research, wrist based devices, in particular, were well tolerated and typically worn without resistance, making them a feasible option for long term use.

________________________________________

Small studies, big potential

Although the data covered just 246 participants, 170 of whom were autistic, the consistency of results across research studies demonstrated a credible ‘proof of concept’. In other words, despite small sample sizes and varied study designs, researchers repeatedly saw similar patterns of early physiological change and reliable detection across different settings.

The implications for clinical practice are significant. Even a short advance window of 60 to 180 seconds allows carers to anticipate distress, adjust the environment, prompt coping strategies or create space for the young person – potentially preventing harm and easing the emotional strain felt by families. For clinicians, such technology also raises the possibility of less reactive, more preventive care models in inpatient and community services.

________________________________________

From prediction to practice

Despite its promise, the field faces challenges. Much of the current evidence comes from controlled environments, meaning researchers must now explore how these devices perform in homes, classrooms and other day to day contexts. To take this further, the authors say longer term studies are needed, designed with the people who will use them, and tested in everyday settings like homes, schools and community services. These would help establish reliability, safety and clinical utility, which is crucial for NHS integration. 

Ethical considerations are also important. Wearable data can be sensitive, capturing intimate details about emotional and behavioural states, raising questions about consent, data security and the risk of over monitoring. As the authors and others point out, ensuring privacy, transparency and meaningful consent will be essential if these tools are to be welcomed and trusted.

________________________________________

A new direction for autism support

For decades, families and practitioners have described behavioural escalations as unpredictable moments that arrive suddenly and leave little room for preparation. But this new research suggests a different possibility: that with the right tools, small physiological shifts can provide important clues before a child reaches crisis point.

By bringing together emerging evidence, the research suggests that AI enabled wearables could help transform how we understand, anticipate and respond to the needs of autistic children. It does not claim that these devices are ready for routine clinical use, but it provides clear evidence that further research is not only justified but urgently needed.

For families long navigating the uncertainty of sudden behavioural storms, this technology may soon offer a much needed forecast. With thoughtful development, careful governance and close collaboration with autistic young people, wrist  and ankle worn sensors may one day help families ‘see the clouds before the storm’.