Table 2.
Features extracted from the EEG
Feature Type | Feature |
---|---|
Frequency Domain | Wavelet energy, total power, peak frequency, spectral edge frequency (80%,90%,95%), power in frequency bands of width 2Hz from 0 to 12Hz with 50% overlap, power in the normalised frequency bands |
Time domain | Curve length, number of maxima and minima, root mean squared amplitude, Hjorth parameters, zero crossings, zero crossings of the 1st and 2nd derivatives, autoregressive modelling error (model order 1-9), skewness, kurtosis, nonlinear energy, variance of the 1st and 2nd derivatives |
Information theory | Shannon entropy, spectral entropy, Fisher information, singular value decomposition entropy |