Electrodermal Activity (EDA) has two components that are measured in microsiemens µS: Tonic and phasic EDA that reflect different aspects of physiological arousal.
Depending on your research goals, you would look into either of. Next to EDA, you can also look into valence, as arousal and valence are indicators of emotional experiences.
The two components
Changes over a longer period of time
Definition: Tonic EDA, also known as Skin Conductance Level (SCL), represents the baseline level of skin conductance over a longer period of time. It reflects the overall state of arousal or alertness of an individual.
Characteristics: It is a slow-moving, steady component and does not fluctuate quickly. Tonic EDA is influenced by general factors like the person’s overall stress level, emotional state, or the time of day. Therefore, try to account for as many variable as possible, but be aware that that is very difficult. This is essentially what makes EDA less popular.
Usage outside games: Tonic EDA is often used to assess chronic stress or long-term changes in arousal. For example, a higher tonic EDA might indicate a person is generally more anxious or stressed over a prolonged period.
Rapid changes towards a stimuli
Definition: Phasic EDA, also known as Skin Conductance Response (SCR), represents rapid changes in skin conductance associated with discrete stimuli, such as a sudden noise or an emotionally charged image.
Characteristics: Phasic EDA is a fast-changing component and reflects momentary increases in skin conductance, typically in response to specific events or stimuli. These responses are short-lived and return to the tonic level quickly after the stimulus has passed.
Usage outside gaming: Phasic EDA is used to assess immediate responses to stimuli. For example, you can track phasic EDA to understand how a person reacts to stressors or emotional events in real-time.
Typical values
This table offers you an overview about values that could be relevant for analysis.
| Measure | Definition | Typical value |
|---|---|---|
| Skin conductance level (SCL) | Tonic level of electrical conductivity of skin | 2-20 µS |
| Change in SCL | Gradual changes in SCL measured at two or more points in time | 1-3 µS |
| Frequency of NS-SCRs | Number of SCRs in absence of identifiable eliciting stimulus | 1-3 per min |
| ER-SCR amplitude | Phasic increase in conductance shortly following stimulus onset | 0.2–1.0 µS |
| ER-SCR latency | Temporal interval between stimulus onset and SCR initiation | 1-3 sec |
| ER-SCR rise time | Temporal interval between SCR initiation and SCR | 1-3 sec |
| ER-SCR half recovery time | Temporal interval between SCR peak and point of 50% recovery of SCR amplitude | 2-10 sec |
| ER-SCR habituation (trial to habituation) | Number of stimulus presentations before two or three trials with no response | 2-8 presentations |
| ER-SCR habituation (slope) | Rate of change of ER-SCR amplitude | 0.01.0.05 µS per trail |
Legend: Key: ER, event-related; NS, nonspecific; SCR, skin conductance response. Reference: Dawson et al. (2007).
Arousal and Valence – understanding emotional experiences
Additionally, if you are interested in specific emotions, you should then observe arousal with valence.
Definition & Characteristics: Valence refers to the intrinsic attractiveness (positive valence) or averseness (negative valence) of an event, object, or situation. It is essentially the “pleasantness” or “unpleasantness” of an emotion. The arousal level “high” (activation) or “low” (deactivation) represents the intensity or energy level of the emotional experience.
Usage: Its usage is universal and usually this matrix reproduced from the original paper by L. Feldman Barrett and J.A. Russell, 1998, is how to visualize your results. Valence captures whether an emotion feels good or bad, while arousal captures how energetically it is felt. Together, they help define the nature and intensity of emotions, allowing for a more nuanced understanding of emotional experiences.
Usage in games: Use a similar matrix, for example after an experiment, and let your players mark their felt emotional experience.
Analysing EDA signals
Depending on your resources, study goals, experience and time, there are different ways of analyzing EDA signals. All of the methods curated here can be performed in an excel sheet without complex equations with the exception of the CDA method requiring Python coding.
-
Decomposition
Depending if you are an advanced data genius, a beginner and resources available (time, tools), you can pick a method around that. The high-pass filtering method is one of the easiest ones to perform and is usually enough. In case you are looking into machine learning and prediction models, you might want to use the CDA method — more reliable, but it requires higher expertise, more time and Python.
High-pass filtering · easiest, usually enough · beginner CDA · reliable, for ML & prediction · advanced, Python -
Normalization: Choosing the Right Method
The choice of normalization method depends on your specific research question, data characteristics, and the need for comparison across subjects or conditions.
Z-score · common in psychophysiological studies Baseline / percent change · event-related analysis Min-max · machine learning or specific rangeZ = (X − μ) / σ # X raw value, μ mean, σ standard deviation -
Visualization: Share your findings
The most rewarding step is visualizing what you have found. The normalization process is already enough to generate valuable insights depending on your research goal.
Plot original signal + separated tonic & phasic Calculate features like SCR amplitude, frequency Plot valence, arousal, tonic & phasic of a game flow
| Step | Method | Research goal | Level | EDA Type |
|---|---|---|---|---|
| 1 – Decomposition | High-pass Filtering | Assess user experience or flow over a longer play session. | Beginner | Tonic |
| 1 – Decomposition | Continuous Decomposition Analysis (CDA) | Assessing long term or rapid stimuli responses | Advanced | Tonic Phasic |
| 2 – Normalization | z-score | Compare data across different individuals or conditions | Beginner | Tonic Phasic |
| 2 – Normalization | Min-Max Normalization | For machine learning applications. | Intermediate | Tonic Phasic |
| 2 – Normalization | Baseline correction | Analyzing changes due to a stimulus or event | Beginner | Phasic |
| 2 – Normalization | Percentage changes | Comparing responses across individuals or conditions with different baseline levels. Alternative to “range changes”. | Beginner | Tonic |
| 2 – Normalization | Range changes | Alternative to “percentage changes”. Comparing responses across individuals or conditions with different baseline levels. | Beginner | Tonic |
Overview of methods and EDA type suggested depending on your research goals. See appendix below for detailed notes on how each method is being used.
Recommended reading
- (2012). Electrodermal Activity (2nd ed.). Springer. This is one of the most comprehensive texts on EDA. It covers the physiological basis, measurement techniques, data processing, and applications of EDA in various fields.
- (2017). Handbook of Psychophysiology. This handbook offers a broad overview of psychophysiological measures, including EDA. It covers the theoretical background, data acquisition, preprocessing, and analytical techniques.
- (2008). Game Usability. Morgan Kaufmann. This one is tailored towards the video games industry and offers you insights into a variety of methods, including psychophysiological ones. I recommend the first edition because it has better user cases.
- (2006). A continuous and objective evaluation of emotional experience with interactive play environments. CHI ’06, 1027–1036. This paper presents a method of modeling user emotional state, based on a user’s physiology, for users interacting with play technologies.
- (2010). Decomposition of skin conductance data by means of nonnegative deconvolution. Psychophysiology, 47(4), 647–658. This paper introduces and explains the Continuous Decomposition Analysis (CDA) method for separating tonic and phasic components of EDA.
- (2009). Time-series analysis for rapid event-related skin conductance responses. J. Neuroscience Methods, 184(2), 224–234. Discusses methods for analyzing phasic EDA responses, particularly in event-related paradigms.
- (2007). The electrodermal system. In Handbook of Psychophysiology (3rd ed.). Cambridge University Press. This chapter provides a thorough overview of the electrodermal system, including measurement, analysis, and interpretation of EDA data.
- — documentation. NeuroKit2 is an open-source Python package for biosignal processing. It provides tools for EDA preprocessing, decomposition, and analysis.
- — MATLAB toolbox. Ledalab is a MATLAB toolbox specifically for EDA analysis, with a focus on Continuous Decomposition Analysis.