Estimation of skin conductance at low frequencies using measurements at higher frequencies for EDA applications

Using low-frequency (LF) alternating current skin conductance (SC) has recently been recommended for electrodermal activity (EDA) measurement, but the method may imply some limitations in sampling rate, which are insufficient for capturing the complete SC waveform. The aim of this study was to assess whether LF SC can be estimated based on skin admittance (SA) measurements at higher frequencies allowing higher sampling rates. SA measurements from 1 Hz to 70 kHz were gathered from 20 healthy human participants, and an interval from 500 Hz to 10 kHz was used to fit a Cole model to the measured SA by means of the nonlinear least squares method. The LF extrapolation of this fit was used to estimate the LF SC at 1, 10, 22 and 30 Hz. The method produced an overestimation of SC by approximately 20%, and the variation in LF SC was preserved by approximately 95%. The overestimation is most likely due to different frequency dependence behavior (dispersion) of SC at the lowest frequencies, which is not accounted for by a single dispersion model. In conclusion, the SA method using high frequency is unsuitable for estimation of the LF SC level, but can probably be used in EDA measurements, which are scored based on the variations in SC. (paper)