Signal Theory Applied to the Analysis of Electroencephalograms

Considered as a special problem in signal analysis, the representation of electroencephalograms (EEG's) involves the problem of determining a unique number for each degree of freedom. A method of representation is usually chosen because of the convenience of making the empirical measurements. One particular method of representation is discussed with emphasis on the problem of making the desired measurements using electronic equipment. The statistical characteristics of EEG signals do not necessarily include all the significant information. This suggests that the goal of EEG signal analysis should be to represent as uniquely as possible the intervals of interest so that the significant information is not discarded in an averaging process.