Procedure for evaluating an EEG in anesthesia or intensive care by means of a computer, which performs the following steps: - determining parameters of the time interval and / or frequencies of measured EEG curves of an EEG detected with a unit (12) for an EEG channel with a multi-electrode arrangement (26), - use the parameters determined in multivariate classification functions and perform an automatic phase division of the EEG in anesthesia or intensive care, - carry out a first stage of analysis of measured EEG curves, in which curve plots are tested to determine scattering signal components, the scattering signal components being those signal components that influence the overall curve plot, but are atypical for an EEG in anesthesia or intensive care, forming such atypical signal components, which are to be determined, of the amount of biological signals as characteristic curves and non-EEG artifacts in anesthesia or intensive care, containing the biological signals of characteristic non-EEG curves in anesthesia or intensive care traced typical epilepsy curves in the form of seizure potentials, - carry out a second analysis stage, which is carried out when there are tests of such signal components that are scattered, performing an artifact analysis, to find out if in the case of signal components that are scattered it could be biological signals or artifacts, - in which in the case of the Artifact analysis using signals detected by artifact sensors, which are configured as EEG electrode strain sensors, checks for the existence of artifacts, when shape modifications are detected by strain sensors, and verifies the existence of biological signals of non-EEG characteristic curves in anesthesia or intensive care, when shape modifications are not detected by deformation sensors. - represent the EEG phase in anesthesia or intensive care, when no dispersing signal components are detected, - do not represent the EEG phase in anesthesia or intensive care, when artifacts
