# Electrodermal activity
The following algorithms are based on EDA signals that are acquried by the movisens sensors edaMove, EdaMove 3, or EdaMove 4.
The following figure shows the signal waveform of a typical skin conductance response (SCR) and the secondary parameters that can be calculated.
Notes for the detection skin conductance responses:
- SCR are detected automatically, making the specification of an event marker neither necessary nor possible.
- The default minimal rise time for the detection of SCR is 0.05uS/s.
- The default minimal amplitude for the detection of SCR is 0.1uS.
- The default maximal rise tome for the detection of SCR is 0.9s.
- Detection Parameters are based on Dawson 2000 (opens new window) and Boucsein 2012 (opens new window)
In more detail the detection of SCR works as follows (extension of the trough-to-peak method, able to detect super-imposed SCR):
- The turning points of the EDA signal are detected
- The local minima and maxima of the EDA signal are determined
- The following cases are distinguished
- Minimum and maximum were found: The detected SCR is a non-super-imposed SCR, if the last half-value point could be determined. If this is not the case, the SCR must be treated as type 2.
- Turning point and turning point were found: The values of the turning points are stored for type 3 treatment.
- Minimum and turn point found: The values of the minimum and turn point are saved.
- Turning point and maximum found: If the last minimum is present and the last turning point is the same as the current first, this is a type 3 SCR.
- If none of the criteria is met, the stored values are deleted.
- SCR are further handled based on their type
- SCR type 1: In case of a non-superimposed SCR, the amplitude as well as the rise time of the SCR are determined by the calculated values. Then the point in time of the half-life is determined. If this cannot be determined, the SCR type 2 treatment comes into effect.
- SCR type 2 treatment: If a half-life of a previous SCR could not be determined because half the amplitude level was not reached, a mark is set. If this mark is active for the next continuous SCR detection and a non-super-imposed SCR is detected, the amplitude of the next SCR is adjusted. For this purpose, an exponential function is determined based on the descent of the last SCR, which further describes the course of the descent.
- Treatment of the SCR type 3: Type 3 SCRs are characterised by the fact that there are at least two turning points between the low and high points. In order to annotate these individually, the point of minimum slope is determined between the turning points. This is done via the second derivative of the filtered EDA signal.
- A plausibility check is performed before the values found are saved. For the super-imposed SCRs of type 3, the minimum rise time and the minimum amplitude were set to 0.5s and 0.05uS correspondingly.
- If the values are plausible, amplitude, rise time, half-life and the product of amplitude and rise time are stored.
Literature:
- Dawson, Michael E., Anne M. Schell, and Diane L. Filion. "The electrodermal system." Handbook of psychophysiology 2 (2007): 200-223. (opens new window)
- Boucsein, Wolfram. Electrodermal activity. Springer Science & Business Media, 2012. (opens new window)
If you want to process the EDA data with a third party program you can
- use the output ReportEdaText
- convert the entire unisens dataset (opens new window) to have csv data files, e.g. with the unisensMatlabTools function convertBin2Csv (opens new window)
# Skin conductance level (EdaScl)
This parameter calculates the skin conductance level (SCL) by applying a low pass filter (0,1 Hz) to the raw EDA signal. It is used for the EdaSclMean parameter. The unit is [µS].
# Mean Skin conductance level (EdaSclMean)
This parameter calculates the mean skin conductance level (SCL) in an adjustable output interval. This parameter is based on EdaScl. The unit is [µS].
movisensLIVE: EdaSclMean can be calculated live on the sensor.
# Number of SCR (EdaScrCount)
This parameter calculates the number of detected skin conductance responses (SCR) in an adjustable output interval.
# SCR amplitudes (EdaScrAmplitudes)
This parameter calculates the time points and amplitudes of each detected skin conductance responses (SCR). The unit is [µS].
# SCR rise times (EdaScrRiseTimes)
This parameter calculates the time points and rise times of each detected skin conductance responses (SCR). The unit is [s].
# SCR half recovery times (EdaScrHalfRecoveryTimes)
This parameter calculates the time points and the recovery times (to the half amplitude) of each detected skin conductance responses (SCR). The unit is [s].
# SCR Energies (EdaScrEnergies)
This parameter calculates the time points and the energies (half product of amplitude and rise time) of each detected skin conductance responses (SCR). The unit is [µSs].
# Mean of SCR Amplitudes(EdaScrAmplitudesMean)
This parameter calculates the mean amplitude of detected skin conductance responses (SCR) in an adjustable output interval. The unit is [µS].
# Mean of SCR rise times(EdaScrRiseTimesMean)
This parameter calculates the mean rise time of detected skin conductance responses (SCR) in an adjustable output interval. The unit is [s].
# Mean of SCR recovery times (EdaScrHalfRecoveryTimesMean)
This parameter calculates the mean recovery time of detected skin conductance responses (SCR) in an adjustable output interval. The unit is [s].
# Mean of SCR energy (EdaScrEnergiesMean)
This parameter calculates the mean energy of detected skin conductance responses (SCR) in an adjustable output interval. The unit [µSs].
# Mean Arousal (EdaArousalMean)
Arousal is a synthetic signal generated from secondary parameters of detected skin conductance responses (SCRs). A special slew limiter is used that reacts faster on rising signals than on falling signals. The mean arousal is calculated in an adjustable output interval. Please select an output interval of 30 seconds or larger for this parameter.
# EDA text file export (ReportEdaText)
This generates the text file ‘Eda.txt’ that contains the time series of the raw eda signal. The values are comma separated. The unit is [uS].