TL;DR
This systematic review identifies and appraises current smart devices and wearables for detecting and monitoring mental health conditions like depression, anxiety, and stress. It highlights that while average heart rate is commonly used, Heart Rate Variability (HRV) and electrodermal activity (EDA) show more promise for accurate detection, with EEG also showing potential for depression.
What Problem Does This Paper Address?
Mental health conditions like anxiety, depression, and stress are increasingly prevalent globally, posing significant health and economic burdens. Early detection and management are crucial, but objective, continuous monitoring outside of clinical settings remains a challenge. Traditional methods often rely on self-reporting, which can be subjective and prone to bias. This paper addresses the need for accessible, portable, and accurate technologies to monitor these conditions in real-time, aiming to improve early intervention and management.
Methods
This paper presents a systematic review of existing literature. The researchers searched four major databases (MEDLINE, CINAHL, Cochrane Central, and PsycINFO) using a combination of primary search terms related to wearable and smart devices and secondary terms for mental health conditions (depression, anxiety, stress). They included peer-reviewed journal publications with full text, regardless of study design or country, focusing on modern technology published from 2010 onwards. Studies were excluded if they were not in English, were editorials, news releases, research highlights, letters, review articles, commentaries, or technical papers. The review critically appraised the selected articles using the Joanna Briggs Institute Critical Appraisal Checklist.
Datasets
Not discussed in the paper. The paper focuses on a systematic review of existing studies, not on the creation or use of new datasets for training or evaluation.
Key Insights and Ideas
- HRV is King for Stress and Anxiety: Heart Rate Variability (HRV) consistently emerged as the most promising physiological metric for detecting both stress and anxiety across numerous studies.
- Average HR is a Poor Substitute: Many commercially available smart devices rely on average heart rate, which the review found to be significantly less accurate for detecting stress and anxiety compared to HRV.
- EEG Shows Promise for Depression: While not yet widely available in commercial devices, specific EEG signatures show strong potential for detecting depression.
- EDA: Accurate but Questionable Reliability: Electrodermal activity (EDA) demonstrated high accuracy for stress detection in some studies, but concerns about its reliability due to factors like motion artifacts and varying electrode pressure were noted.
- Multimodal Sensing is the Future: Combining multiple physiological signals (e.g., EEG and HRV, or EDA and HR) appears to offer superior accuracy in detecting stress compared to single-metric approaches.
- Wearables are Becoming More Tolerable: The review highlights a trend towards more compact and wearable devices that are better tolerated by users, increasing the feasibility of long-term monitoring.
Key Results
The review identified 21 studies that met the inclusion criteria. Key findings include:
- Stress and Anxiety Detection:
- HRV was the primary physiological marker used in 10 out of 15 studies that detected stress.
- Combined EEG and HRV showed higher accuracy (87.5%) in detecting stress than EEG alone (77.9%) or HRV alone (75.0%).
- EDA demonstrated high accuracy for stress detection (e.g., 85.3% classification accuracy in one study).
- Respiratory rate and variability were also explored for stress detection, with mixed results.
- Depression Detection:
- EEG was used in two studies for detecting depressive symptoms, with one study achieving 99.1% accuracy for mild depression using k-nearest neighbour classifiers.
- Physical activity measured by accelerometers and smartwatches, combined with machine learning, showed promise for detecting depression (e.g., 96% accuracy in one study).
- Functional Near-Infrared Spectroscopy (fNIRS) also showed high accuracy (92.6%) in classifying depression.
Limitations and Future Work
- Device Availability: While many wearable technologies show promise in research, the review notes that devices specifically designed for depression detection using EEG are not currently available on the market.
- Validation: Many wearable devices have not been formally validated for research or clinical use, and only a small percentage of those reviewed had undergone such validation.
- Search Scope: The authors acknowledge that devices not yielding from their specific search terms might have been missed, potentially impacting the comprehensiveness of the conclusions.
- Future Directions:
- Further research is needed to assess the applicability of dual-device approaches (e.g., EEG + HRV) for long-term chronic stress monitoring.
- Investigating the combination of accelerometers and EEG could yield promising results for depression diagnosis.
- More research is needed on remote measurement technologies that can objectively assess factors like sleep, sociability, and cognitive function, which are indicators of depression.
- Validating wearable technologies and standardizing parameter recording for smart devices is crucial for their integration into healthcare.
- Multiparametric devices are likely to offer the greatest potential for accurate detection of mental health conditions and stress.
Summary updated