How does it work?
Each device combines several sensors that monitor different physiological parameters typically affected by stress: heart rate, skin conductivity and temperature, muscle tension, facial expressions, tone of voice, etc. These various sets of data are merged and compared to reference values in order to estimate stress levels.
Advantages?
CEA-Leti designs its devices in collaboration with psychology and biomedical research laboratories. This enables researchers to define the relevant physiological parameters on a case-by-case basis and then select the best sensors to measure them.
Another strong point is that the interpretation of these measurements takes into account the diversity of stress reactions from one individual to another as well as confounding factors likely to distort measurements (e.g., physical activity).
These innovative designs are for wearable devices that make it possible to provide users with continuous stress monitoring. This technology is protected by seven patents.
Our research areas:
- Physiological sensor characterization and artifact detection
- Merging heterogeneous data
- Machine learning and artificial intelligence methods
- Inter-individual variability in stress responses
Applications?
- Stress monitoring for cases of hypertension and diabetes
- Early detection of relapses for bipolar disorder
- Accident prevention in high-risk work environments
If you want to monitor stress in order to track a pathology or prevent accidents, you cannot rely on estimates from connected watches. We're aiming to develop medical-grade measurements that are validated by experts in the field.” Christelle Godin, signal processing research engineer at CEA-Leti