An overview is presented of an expert system approach to the automatic interpretation of recordings of the electrical activity of the brain (electroencephalogram, or EEG). The method consists of a quantification module and a rule-based expert system. Quantification is done by means of autoregressive spectral estimation. It was found that the rule-based system correctly classified 93% of EEGs taken from renal patients, and a 70% agree-ment was found between the system and an electroencephalographer in an experiment aimed at determining the depth of sleep in normal subjects. An indication of how this approach may be used for spectral pattern recognition concludes this paper.