Quantitative EEG: I. Techniques and problems of frequency analysis and topographic mapping.

Abstract

Quantitative EEG techniques include frequency analysis (spectral analysis), significance probability mapping, and other analytic techniques. Each can be done on spontaneous EEG in various states or in conjunction with sensory stimulation. Several types of displays are available, including topographic mapping of scalp electrical activity. Assessment of normality in these records must take into account age, gender, state of alertness, medications, and other factors. Substantial statistical issues are critical in these assessments and must be thoroughly understood by all users. Other problems can easily mislead the interpretations of these tests, sometimes in subtle ways. References are often active. Traditional EEG artifacts can appear in surprising ways, and new artifacts can be caused by computer processing and display format. Important technical choices must be made in recording quantitative EEG, and the correct choices are not clearly known. These choices include references, number of channels, epoch length, number of epochs acquired, and artifact rejection criteria. This review summarizes a variety of the techniques commonly used. Advantages of particular methods are contrasted. Problems with these techniques are discussed at length, with emphasis on the difficulties and choices facing users of typical commercial quantitative EEG machines. As quantitative EEG techniques come into some clinical use, issues of nomenclature, technique, normality, and problems will become widely understood. For now, clinicians should respect the problems inherent in these techniques. Quantitative EEG tests should only be interpreted along with the traditional paper EEG tracing that represents the raw data on which the quantitative analysis was performed. A thorough familiarity with traditional EEG is a prerequisite to understanding the meaning of the quantitative EEG results.