1. Home
  2. / Science and Technology
  3. / Now it’s a reality: researchers at Unicamp develop AI capable of identifying anxiety with 80% accuracy directly on a smartwatch, paving the way for continuous mental health monitoring with low energy consumption.
Reading time 5 min of reading Comments 0 comments

Now it’s a reality: researchers at Unicamp develop AI capable of identifying anxiety with 80% accuracy directly on a smartwatch, paving the way for continuous mental health monitoring with low energy consumption.

Written by Hilton Libório
Published on 13/06/2026 at 16:41
Be the first to react!
React to this article

Researchers from Unicamp create AI to detect anxiety in smartwatches with 80% accuracy, expanding advances in digital health. 

Wearable technology has just reached a new level in Brazil. Researchers from Unicamp have developed an AI capable of identifying signs of anxiety with more than 80% accuracy using data captured by a smartwatch. The project is conducted by the Viva Bem Applied Research Center: artificial intelligence for health and well-being, funded by FAPESP and Samsung.

The results were presented by Professor Anderson Rocha during FAPESP Week London, held from June 2 to 4. The proposal is to transform devices used daily into tools capable of continuously monitoring emotional and physiological changes, helping to prevent health problems.

According to an article by Agência Fapesp, published by Elton Alisson on June 8, 2026, more than detecting anxiety episodes, the initiative paves the way for future applications involving hypertension, diabetes, Parkinson’s, and fall prevention in the elderly.

Researchers from Unicamp advance in the use of AI to monitor anxiety

The project developed by the researchers aims to use body signals captured by a smartwatch to identify moments of anxiety even before the person consciously realizes what is happening.

The AI analyzes different physiological information collected throughout the day and learns to recognize individual behavior patterns. This capability allows creating a kind of personalized profile for each user.

According to Anderson Rocha, coordinator of the research center, the first version of the technology has already been published scientifically. Currently, new improvements are being evaluated by Samsung, the company responsible for the hardware used in the tests.

How the smartwatch collects data to train the AI

The system mainly uses two types of information captured by the smartwatch:

  • Electrocardiogram (ECG);
  • Accelerometry data from arm movements.

These signals are processed by the AI to form a unique “data signature” for each individual. From this set of information, the algorithms can identify changes that may be related to anxiety states.

The differential lies precisely in the combination of these data. Instead of analyzing only heartbeats, the technology considers multiple indicators simultaneously.

This approach tends to increase the accuracy of the analysis and reduce misinterpretations.

Controlled tests help researchers identify emotional patterns

To teach the algorithms to differentiate emotional states, researchers created specific stress induction protocols.

One of the experiments requires participants to perform complex mental calculations in just 30 seconds. Among the challenges applied is solving multiplications like 309 by 17 while a countdown appears on the smartwatch screen.

The situation generates controlled psychological pressure and allows observing how the body reacts to stress.

During the process, the AI records physiological changes and learns to associate these signals with the moments of anxiety observed in the participants.

It was based on this training that the system achieved more than 80% accuracy in the results disclosed so far.

AI does not replace doctors, but can enhance prevention

Researchers emphasize that the technology was not created to provide medical diagnoses.

The main goal is to function as an early warning tool. If the AI detects recurring episodes of anxiety, the system may recommend that the user seek professional evaluation.

In practice, the smartwatch would act as a permanent health observer, offering additional information that can help in decision-making.

Among the expected benefits are:

  • Early identification of emotional changes;
  • Encouragement to seek specialized support;
  • Continuous health monitoring;
  • Greater awareness of one’s own body signals.

This strategy follows a growing trend of data-driven preventive medicine.

Unicamp expands research to other health conditions

Anxiety monitoring represents just one of the fronts explored by researchers.

The team is also working on solutions aimed at other conditions that can be identified through physiological signals captured by wearable devices.

Among the studied topics are:

  • High blood pressure;
  • Diabetes;
  • Parkinson’s disease;
  • Risk of falls in the elderly;
  • Cardiovascular changes.

The expectation is that AI will be able to recognize early signs of these conditions, allowing for quicker interventions and increasing the chances of better outcomes for patients.

The ultimate goal is to use the data collected by the smartwatch to improve people’s quality of life through the early detection of health problems.

Increase in anxiety cases boosts the relevance of technology

The advancement of research occurs at a time when mental health is gaining more and more attention worldwide.

Anxiety is part of the body’s natural responses, but when it becomes frequent or intense, it can affect well-being, relationships, and professional performance.

Among the most common symptoms are:

  • Excessive worry;
  • Insomnia;
  • Irritability;
  • Difficulty concentrating;
  • Increased heart rate.

For this reason, tools capable of identifying early signs can play an important role in the prevention and monitoring of these conditions.

In this scenario, the combination of AI and smartwatches emerges as a promising alternative to expand mental health monitoring.

Regulatory approval will be a decisive step to bring innovation to the public

Despite the positive results, the technology is still undergoing processes of improvement and validation.

Researchers continue to refine the algorithms and expand the tests to increase the reliability of the analyses. Only after this stage will it be possible to advance to more comprehensive studies with real users.

According to Anderson Rocha, when the solution reaches sufficient maturity, authorizations will be requested from regulatory bodies, including the National Health Surveillance Agency (Anvisa).

This process is essential to ensure that the system meets the safety and quality requirements demanded for health-related applications.

What this advancement reveals about the future of digital health

The work developed by Unicamp researchers demonstrates how the integration between AI, biometric sensors, and smartwatches can transform the way people monitor their own health.

By achieving more than 80% accuracy in identifying anxiety, the research shows that devices used daily can become important allies in preventive medicine. Although there are still regulatory and technical steps ahead, the project already indicates a future where wearable technologies will be able to detect early signs of various conditions, helping users and professionals to act more quickly and efficiently.

Sign up
Notify of
guest
0 Comments
most recent
older Most voted
Hilton Libório

Hilton Fonseca Liborio is a writer with experience in digital content production and SEO skills. He specializes in creating optimized content for diverse audiences and platforms, aiming to combine quality, relevance, and results. His areas of expertise include the Automotive Industry, Technology, Careers, Renewable Energies, Mining, and other topics.

Share in apps
0
I'd love to hear your opinion, please comment.x