The system was then evaluated in a diverse international cohort to assess whether the wearable’s physiological signals were associated with stress, anxiety, and depression. These findings, published in Communications Medicine, demonstrate the potential of wearable sensors for non-invasive mental health monitoring and provide a foundation for assessing psychological wellbeing.
Limitations of Traditional Psychological Assessments
Traditional psychological stress assessments rely heavily on self-reported questionnaires and physiological measures, such as heart rate variability. These methods can be affected by reporting bias and environmental conditions, reducing measurement reliability.
To address these limitations, researchers developed a wearable optical sensor that combines laser Doppler flowmetry and fluorescence spectroscopy. Laser Doppler flowmetry measures microvascular blood flow, while fluorescence spectroscopy detects metabolic coenzymes that reflect cellular activity. Together, these methodologies provide objective measurements of vascular and metabolic responses associated with psychological stress.
Saving this for later? Download a free PDF now.
Design and Methodology: The Optical Sensor
The study involved 132 adults (aged 18–94 years) from 19 countries who wore a non-invasive device to measure fingertip blood flow and tissue metabolism during 15-minute sessions while resting supine. The sensors continuously recorded their physiological signals, including blood perfusion, skin temperature, and tissue metabolic activity.
Wavelet analysis was employed to separate oscillations associated with endothelial, neurogenic, myogenic, and cardiac regulation. All participants also completed the 21-item Depression, Anxiety, and Stress Scale (DASS-21) questionnaire. This enabled the comparison of physiological measurements with standardized psychological assessments and created a comprehensive, well-informed dataset. The resulting multimodal dataset was then used to evaluate machine learning models for detecting psychological stress.
Outcomes from Physiological Data on Psychological Distress
The multimodal dataset demonstrated that combined physiological and demographic data contain signals associated with questionnaire-reported psychological distress. However, the model’s performance was insufficient to reliably establish individual assessment.
Among the machine learning models evaluated, the Light Gradient Boosting Machine (LightGBM) achieved the best performance. Using the 10 most important features, the model reached a receiver operating characteristic area under the curve (AUC) of 0.72 and a precision-recall AUC of 0.89 during cross-validation.
More than 27% of participants reported symptoms of stress, anxiety, depression, or a combination of these on the DASS-21 questionnaire. These results demonstrate that integrating microvascular and metabolic measurements may help to distinguish individuals with psychological distress from those without it.
To explain the model's predictions, researchers applied explainable artificial intelligence (XAI) methods. Body mass index (BMI) emerged as the strongest predictor, followed by age, sex, and resting heart rate. Notably, female participants reported higher levels of psychological distress than males, and middle-aged individuals had the highest mean DASS-21 scores.
The optical measurements contributed to the model's performance. Features derived from laser Doppler flowmetry and fluorescence spectroscopy captured changes in microcirculation and tissue metabolism associated with psychological distress. However, performance declined when wearable-derived features were analyzed without demographic variables.
Participants experiencing psychological distress often showed greater fluctuations in blood flow and metabolic signals compared to the healthy group, whose physiological measurements remained stable. These markers provided additional physiological information associated with DASS-21 classifications, supporting further investigation of the use of wearable optical sensors for mental health assessment.
Applications for Continuous Mental Health Monitoring
These wearable optical sensors could have significant implications for continuous mental health monitoring outside clinical settings. Unlike traditional psychiatric assessments, which rely on scheduled evaluations and self-reported symptoms, these devices provide objective, real-time measurements of physiological changes associated with psychological distress during a recording session.
However, further studies are needed to determine whether the approach can support continuous monitoring during everyday activities.
In the future, integrating laser Doppler flowmetry and fluorescence spectroscopy into smartwatches or medical wearables could support earlier detection of stress-related conditions and enable continuous monitoring of treatment responses. Such technology could be developed to support mental health care by identifying changes in physiological markers before symptoms become severe, thereby allowing for earlier clinical intervention and more personalized care.
Conclusion and Future Directions
This study demonstrates that optical measurements of microvascular blood flow and tissue metabolism can provide objective markers of psychological distress. Combined with machine learning, these signals may offer a promising complementary source of information for mental health assessments and support the development of non-invasive, wearable monitoring systems.
Future work should focus on further expanding the dataset, evaluating more diverse populations, and applying deep learning models to improve the interpretation of complex physiological signals. Overall, these advancements could support more reliable, personalized, and continuous mental health assessment in clinical and community settings.
Journal References
Nguyen, M.N., et al. (2026). A wearable device dataset for mental health assessment using laser Doppler flowmetry and fluorescence spectroscopy sensors. Communications Medicine. https://www.nature.com/articles/s43856-026-01766-5.
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.