In order to determine whether the clinically used frequency bands of the EEG can also be obtained by a mathematical system we did a factor analysis with 480 EEG recordings, 5 minutes each, in 60 healthy male volunteers. A power spectrum analysis was done and 57 frequency bands between 1.5 and 30.0 Hz in a half Hz steps were calculated. The factor structure obtained made the following frequency bands (Hz) reasonable: deltaF = 1.5 - 6.0, thetaF = 6.0 - 8.5, alpha1F = 8.5 - 10.5, alpha2F = 10.5 - 12.5, beta1F = 12.5 - 18.5, beta2F = 18.2 - 21.0, beta3F = 21.0 - 30.0. Except for alpha1F all other 6 frequency bands were represented by one general factor with the complexity 1. The variance of the alpha1F band is explained by several of the 6 factors. The clinically known and the by factor analysis obtained frequency bands in the beta-range are similar. The clinically alpha-band is subdivided into two frequency bands alpha1F and alpha2F by the factor analysis. The clinically known border line between delta- and theta-band of 3.5 Hz cannot be found by factor analysis.