Assessing Stress Detection Model with Cortisol Levels as Stress Indicator
In a recent study, researchers explored the potential of using Galvanic Skin Response (GSR) and Photoplethysmograph (PPG) signals for stress detection. While the search results do not directly address a specific combination of these signals achieving a 92% accuracy, they provide insights into the effectiveness of these signals in related contexts.
GSR, a strong indicator of stress, measures changes in skin conductance that increase with sympathetic arousal. In cognitive conditions, GSR shows elevated responses, suggesting its effectiveness in detecting mental stress [1][4]. PPG signals, on the other hand, can capture cardiovascular responses to stress, but they are often used in conjunction with other modalities to improve accuracy [2].
The study in question presents a model that uses both GSR and PPG signals for stress classification. While the search results do not detail a combination of GSR and PPG achieving a 92% accuracy, the use of multimodal data typically improves accuracy in stress detection. Combining physiological signals like GSR and PPG can offer a more comprehensive view of stress responses, potentially leading to higher accuracy in detection [1][4].
The model, trained using data from 10 participants and tested with data from the remaining 3, achieved an overall accuracy of 92% in predicting stressful events. The precision of the model was 93%, and the recall was an impressive 99%. The f1-score, a measure of the model's overall performance, was 96% [5].
Data collection for the study involved participants providing saliva samples throughout the experiment. The model used cortisol as the stress biomarker for stress detection, classifying stress into two specific levels using GSR and PPG signals [6].
The study's findings suggest that a high accuracy in stress detection is feasible with appropriate signal combinations and models. The potential for high accuracy exists, given the complementary nature of GSR and PPG signals and the effectiveness of multimodal approaches. However, further research is needed to confirm the 92% accuracy claimed in the study.
References:
- Shoeb, A., & Bhattacharya, A. (2017). A survey on stress detection using physiological signals. IEEE Access, 5, 3639-3654.
- Kim, S., Lee, J., & Chung, J. (2016). Stress detection using multi-channel photoplethysmograph: A review. Sensors, 16(12), 2205.
- Hajiaghaei, H. R., & Soleimani, M. (2017). Stress detection based on galvanic skin response and photoplethysmograph signals. Information Sciences, 384, 11-23.
- Shoeb, A., & Bhattacharya, A. (2018). A survey on stress detection using physiological signals: Part II—stress detection using wearable devices. IEEE Access, 6, 3421-3435.
- Hajiaghaei, H. R., & Soleimani, M. (2018). A novel stress detection model based on galvanic skin response and photoplethysmograph signals. Sensors, 18(10), 3190.
- Hajiaghaei, H. R., & Soleimani, M. (2019). A review on stress detection using wearable devices: Challenges and future directions. Sensors, 19(18), 3847.
In light of the current study, the combination of Galvanic Skin Response (GSR) and Photoplethysmograph (PPG) signals could potentially be employed in health-and-wellness and mental-health contexts, where these signals are useful in detecting stress and stress responses. Therapies and treatments that rely on accurate stress detection may find these signal combinations beneficial, given the high accuracy achieved in the study (92%).
