A statistical approach to computerized EEG: preliminary data on control subjects and epileptic patients

Factor Analysis represents an important method of extracting salient features from EEG data condensing functional as well as spatial informations in a few factors and so reducing redundancy of multi-channels computerized EEG data. A 16-channel computerized frequency analysis of background brain electrical activity during 3 functional conditions (eyes closed, eyes open and hyperventilation) were performed in 2 groups, 50 healthy subjects and 14 epileptic patients with generalized seizures. The differences in each frequency band between the 2 groups have been assessed by means of factor scores. Factor scores differences between control and epileptic patients have been found in theta and alpha frequency bands. These interictal EEG abnormalities were more evident in epileptic patients. This preliminary study shows the confirmatory capability of Factor Analysis and its usefulness in managing and comparing computerized EEG data.