Research Shows Stress Level Can Be Detected Using Galvanic Skin Response
Like so many things in life, stress is a double-edged sword. As an adaptive mechanism, stress keeps us alert and motivated, allowing us to achieve goals and stay safe from danger. However, when stress turns from an acute episode to a chronic condition, our health begins to suffer. Stress has been associated with a wide variety of diseases and disorders, including immune deficiencies, diabetes, neurodegenerative disorders, heart disease, and depression.
In the busyness of modern-day living, modern-day solutions can be enormously helpful in recognizing our stress levels. If stress can be accurately detected, we have a better chance of mitigating chronic stress and warding off its associated ailments. Research is increasingly uncovering that galvanic skin response is one such promising approach.
The stress response
The body reacts to good or bad changes in the environment by initiating a stress response. This stress response is reflected in a wide range of physical symptoms. These symptoms are governed by the autonomic nervous system (ANS), which regulates largely involuntary bodily functions.1
A branch of the ANS called the sympathetic (or “fight or flight”) nervous system is mobilized in times of stress, leading to a cascade of physical effects such as increased heart rate, a spike in blood pressure, quickening respiration, and increased perspiration.1 As we’ll explore next, sweating is the symptom of stress that is capitalized on by galvanic skin response sensors.
Measuring stress with galvanic skin response (GSR)
When the sympathetic nervous system activates our sweat glands in the midst of a stressor, the electrical characteristics of our skin change. These changes are seen as increased skin conductance, also known as an electrodermal activity. That is to say, our skin momentarily becomes a better conductor of electricity in response to an emotionally-arousing stimulus.2
By measuring the increase in electrical conductance of the skin, galvanic skin response sensors allow for an objective way to measure our stress levels. These sensors are capable of recording data continuously and in a variety of environmental conditions. From here, the GSR data can be further analyzed with the help of cutting-edge models. For instance, researchers can use machine learning algorithms to build stress-detection models with a training dataset, then accurately detect stress levels with this model using a testing dataset. It’s then possible to find the relationship between GSR signal variations and external events or factors that are associated with stress, with the ultimate aim of mitigating their impact.
In recent years, economic and unobtrusive devices capable of collecting continuous GSR data have emerged. These devices, including smartwatches and wristbands, have ushered in a new body of research investigating GSR’s efficacy in detecting stress levels, and how these levels vary with daily life events like driving.
What the research is saying
A study from 2018 sought to classify driving stress using data collected from a wearable GSR device. GSR measurements were made during rural road driving (the low-stress condition) and highway or highway under construction driving (the high-stress condition). The researchers used GSR data from 9 drivers to build a model that would identify driving stress. Then, the accuracy of this model was validated with GSR data collected from a single driver.
The study found an overall classification accuracy of over 83% with the testing dataset, indicating high performance on detecting low vs. high stress, without the need to collect multiple physiological signals.3 The study authors conclude that such a model could be integrated into commercial wearable devices equipped with GSR sensors, allowing real-time stress-level detection and mitigation.
An earlier study from 2013 used GSR data and its associated features to classify the stress levels of 10 subjects. The subjects underwent easy and difficult arithmetic tests—corresponding to light and heavy workloads. Classifiers trained on the GSR data showed a maximal accuracy of 70%, with high variability between subjects.4 In contrast to the previous study, the researchers concluded that GSR should be combined with other measurements to obtain optimal detection performance, especially when detecting stress levels in real-life settings.
GSR with other measurements
When the GSR signal rises, this can indicate arousal and therefore a stressful event. But due to how the sensors are placed, changes from the baseline can also reflect contextual factors such as a rise in temperature, heavy physical work, and movement.5 Since real-world stress frequently involves movement, GSR can be combined with additional sources of information to gain insight into whether the observed patterns are due to stress or something else.
One combination that has been investigated is GSR coupled with an accelerometer. By measuring movement, accelerometers can shed light on how activity levels impact GSR signal peaks, particularly peaks that aren’t caused by stress alone. A 2014 study combined GSR with two accelerometers (placed at the thigh and ankle) to continuously detect stress in 5 subjects. The subjects were required to sit, stand, and walk while undergoing either rest or stressful mental arithmetic tests. The researchers found that the activity information provided by the accelerometers led to an improved stress detection rate at the sitting and standing positions.6
GSR biofeedback for stress
In addition to detecting and classifying stress based on activities, GSR data can also be used to directly regulate stress through the use of biofeedback. The rationale of GSR biofeedback centers on self-awareness: if we can be more conscious of what physical and psychological factors influence our stress levels, we can more easily make changes in our behavior and lifestyle for stress management.
A popular form of GSR biofeedback involves using an auditory tone as feedback that varies with skin conductance measured at the fingertips. Typically, this tone increases in pitch when skin conductance increases in response to stress. When stress levels are lowered with learned relaxation techniques, skin conductance decreases, along with the pitch of the tone. In this way, real-time feedback based on GSR allows us to detect subtle changes in our emotional states, and then these states can be pushed into an optimal direction with relaxation training. GSR biofeedback is commonly done using home devices or with a professional in psychotherapeutic settings.
Stress levels and ZYTO GSR technology
While ZYTO technology doesn’t directly detect stress levels, the galvanic skin response readings which the Hand Cradle gathers are used to create what we call a Dynamic Profile. From this profile, we can observe the responses that fall outside of the baseline range.
Responses to many different Virtual Items can be observed, such as those representing systems and parts of the body, environmental factors, and more. In this way, ZYTO technology goes beyond stress detection and allows us to see how the body responds to a variety of items.
Not only that, but ZYTO also scans Virtual Items that can be used to balance the out-of-range stressors, such as supplements and essential oils. Wellness professionals can then use this information to support overall wellness, which may in turn help an individual to better manage their stress.
1.. Yu, B., M. Funk, et al. “Biofeedback for Everyday Stress Management: A Systematic Review.” Frontiers in ICT 5 (2018).
2. Sharma, M., S. Kacker, & M. Sharma “A Brief Introduction and Review on Galvanic Skin Response.” International Journal of Medical Research Professionals 2, no. 6 (2016): 13-17.
3. Kim, J., J. Park, & J. Park. “Development of a statistical model to classify driving stress levels using galvanic skin responses.” Human Factors and Ergonomics in Manufacturing & Service Industries 30, no. 5 (2020): 321-328.
4. Kurniawan, H., A.V. Maslov, & M. Pechenizkiy. “Stress detection from speech and Galvanic Skin Response signals.” Proceedings of the 26th IEEE International Symposium on Computer-Based Medical Systems (2013): 209-214.
5. Bakker, J., M. Pechenizkiy, & N. Sidorova, “What’s Your Current Stress Level? Detection of Stress Patterns from GSR Sensor Data.” 2011 IEEE 11th International Conference on Data Mining Workshops (2011): 573-580.
6. T.B. Tang, L. W. Yeo, & D. J. H. Lau. “Activity awareness can improve continuous stress detection in galvanic skin response.” SENSORS, 2014 IEEE (2014): 1980-1983.