University of Geneva Study Proves Smartwatches and AI Can Predict Fluctuations in Cognitive and Mental Health

University of Geneva study shows AI and smartwatches predict cognitive and emotional health with 87.5% accuracy, aiding early detection of brain disorders.

By: AXL Media

Published: Mar 10, 2026, 4:19 AM EDT

Source: The information in this article was sourced from Université de Genève

The Rise of Non-Invasive Brain Monitoring

As global populations age, the prevalence of neurological and mental disorders is reaching a critical threshold, with the World Health Organization reporting that over one in three people live with conditions like stroke or Parkinson’s. In response to this public health challenge, researchers at the University of Geneva have explored a proactive prevention strategy utilizing the technology already present on millions of wrists. By monitoring 88 volunteers between the ages of 45 and 77, the team demonstrated that smartphones and smartwatches can gather a "passive" stream of biological and environmental data capable of forecasting changes in brain health. This approach offers a continuous, non-invasive alternative to traditional clinical assessments, which often only capture a snapshot of a patient's health during an office visit.

Capturing the Daily Rhythms of Cognitive Health

The study tracked 21 distinct indicators over a ten-month period, collecting data on physical activity, sleep patterns, and heart rate without requiring any changes to the participants' daily habits. This "passive" data collection was supplemented by "active" data, where volunteers completed emotional questionnaires and cognitive performance tests every three months. According to Igor Matias, the lead author of the study and a doctoral assistant at the Geneva School of Economics and Management, the goal was to determine if human brain health—which fluctuates naturally based on lifestyle and environment—could be mapped with enough precision to identify the early warning signs of more serious abnormalities.

Artificial Intelligence as a Predictive Engine

The core of the UNIGE project relied on a custom-developed artificial intelligence model designed to find correlations within the massive dataset. When the AI’s predictions were compared against the participants' actual test results, the researchers found a remarkably low average error rate of 12.5%. This success indicates that machine learning algorithms can successfully interpret the "noise" of daily life—such as fluctuations in sleep quality or heart rate—to provide an accurate forecast of an individual's emotional and cognitive state. The precision of these models marks a major step forward in the use of wearable technology for clinical-grade health diagnostics.