July 16th, 2026

4 min read

Autism care often relies on observation and caregiver reports, making it difficult to identify anxiety, stress or behavioral escalation before a crisis occurs.

To help clinicians intervene earlier and more proactively with young autism patients, NTT DATA Europe & Latam (NTT DATA Spain) and Fundació MútuaTerrassa collaborated on AI4ASD, a project that uses wearable sensors, cloud analytics, and AI to identify physiological changes suggesting that something is wrong before outward signs of escalation become visible.

NTT DATA Spain developed the project’s secure AI and analytics platform ensuring transparency, accountability and trust while exemplifying how responsible AI can help healthcare organizations adopt predictive technologies.

The rise of AI in healthcare

Autism Spectrum Disorder (ASD) affects millions of people worldwide and presents unique challenges such as anxiety and stress for patients, families, caregivers, and healthcare professionals. Clinicians mainly count on their expertise as physicians, behavioral observation, caregiver feedback, and subjective assessments to understand a patient’s condition. While these approaches remain essential, they can make it difficult to detect signs of distress before a behavioral crisis emerges.

As a result, healthcare teams are frequently placed in a reactive position of responding once a patient begins shouting, crying, or exhibiting signs of escalation rather than intervening beforehand. This entails an increased stress for families and caregivers, and reduces the ability to personalize care based on individual physiologic patterns.

The challenge is becoming increasingly urgent as autism diagnoses continue to rise globally. Healthcare systems must find ways to support more patients while maintaining quality of care and controlling costs. Recognizing this need, Hospital Universitari MútuaTerrassa partnered with NTT DATA Spain to explore whether artificial intelligence and physiological monitoring could provide clinicians with actionable patient outcomes and help shift autism care from reactive intervention to proactive support while upholding the ethical standards of the healthcare industry.

Using AI to predict Autism-related behavioral crises

AI4ASD is a collaborative initiative between NTT DATA Spain and Fundació MútuaTerrassa designed to identify physiological signals associated with stress, anxiety, and potential behavioral crises in children with severe autism.

The project uses wearable technology to continuously collect biometric data, including heart rate variability, electrodermal activity (EDA), skin temperature, movement and activity patterns, and sleep-related indicators. The data is securely transmitted to a cloud-based smart health analytics platform developed by NTT DATA Spain, where AI models analyze physiological patterns and identify anomalies that may indicate a patient is becoming distressed.

During a nine-month pilot project involving 30 patients with severe autism, clinicians explored how these insights could help identify potential escalation earlier than traditional observation alone. In some cases, physiological changes appeared before visible behavioral signs emerged, creating opportunities for caregivers to engage patients, understand potential triggers, and potentially prevent situations from escalating further.

According to Jose Aznar, AI4ASD Project Manager, NTT DATA Spain, “So far, the only way clinicians could act was reactively. Something happened, the child started to shout or cry, and then we reacted. This project is about acting the other way around.” The ultimate goal, he says, is to develop responsible AI to provide healthcare professionals with better tools to personalize care, improve patient outcomes, and reduce stress for families and support communities.

Building AI technology that transforms biometric data into actionable insights

The team at NTT DATA Spain worked to design and implement the secure cloud and AI infrastructure that powers AI4ASD. In doing so, it was critical that AI4ASD met all ethical standards to earn the trust of the other stakeholders, including patients, their families, caregivers, and healthcare professionals.

The system works like this: The clinician fits an autistic child with a wearable device, and the platform continuously receives biometric information through the device, transforming large volumes of physiological data into meaningful insights clinicians can use during patient care. Using advanced machine learning techniques, the system learns patterns associated with an individual’s typical physiological state and identifies deviations that may indicate rising stress or behavioral instability.

One of the most promising outcomes from the pilot involved patients with aggressive behavioral profiles. The AI models were able to identify physiological changes suggesting that something was wrong before outward signs of escalation became visible. This allowed clinicians to count on a tool that enables them to engage patients proactively—asking simple questions such as whether they were uncomfortable, hungry, cold, or experiencing anxiety—rather than waiting until a crisis had already begun.

While the project demonstrated encouraging results, Aznar explains that NTT DATA Spain and clinical partners view it as an important first step rather than a finished solution. Additional research, larger patient populations, and broader clinical validation will be required before predictive models can be applied at greater scale.

Advancing ethical AI in healthcare

Aznar adds that AI4ASD also became an important learning experience in the responsible use of AI within healthcare. When the project began, the use of AI in European healthcare settings was still relatively new.

As a result, NTT DATA Spain worked closely with hospital leadership, clinicians, legal teams, and ethics committees to address critical questions around data privacy, transparency, explainability, and trust.

A key requirement of the project was explainability, ensuring that clinicians could understand how AI-generated insights were produced. Rather than operating as a “black box” where the AI doesn’t explain the logic of its insights, the system was designed to support explainable decision-making and maintain physician oversight at every stage. The project also adhered to strict ethical principles that prohibit discriminatory use of AI and prevent AI systems from independently determining access to healthcare services or treatment options.

The future of AI in behavioral healthcare

The AI4ASD pilot demonstrated the potential of combining smart physiological monitoring and AI in an ethical manner to support earlier intervention, but the team sees far broader possibilities ahead including leveraging advances in AI to improve scalability and clinical usability.

Aznar says future versions of the platform could enable patients and caregivers to receive earlier warnings about rising stress levels while remaining at home in familiar environments, supporting more independent living and reducing the need for intensive clinical supervision. Importantly, the long-term vision is to provide healthcare professionals with better visibility into each individual’s unique needs.

By combining human expertise with the responsible use of AI, AI4ASD could help create a future where autism care becomes more personalized, proactive, and accessible for patients, families, and healthcare providers alike.

And beyond its clinical value, AI4ASD reflects NTT’s broader sustainability and innovation vision. The project’s impact was recently recognized with the Inclusive Society Award at the NTT Group Sustainability Conference.