Abstract
As an electroencephalogram (EEG) is capable of reflecting at least part of the functional status of a patient's brain, EEGs interpreted by electroencephalographers (EEGers) are used as a diagnostic aid for diseases that affect the brain. Over the past 10 years, the authors have worked to develop an automatic interpretation technique for awake background EEGs. Automatic integrative interpretation of awake background EEGs in clinical use requires the appropriate selection of EEG segments from a lengthy EEG record, which may contain various artifacts such as eye blinks and change in characteristic according to the patient's vigilance level or open-and-closed eye movement during the recording period. In this study, a method for automatically selecting the EEG segments was newly developed by combining the automatic recognition of dominant rhythm organization, artifacts, vigilance level and open-eye state of the patients. The method was applied to the lengthy EEG records of 30 patients. The EEG segments were appropriately selected by the proposed method, and then subjected to the automatic background EEG interpretation method. The proposed method avoided serious mistakes in interpretation and improved the accuracy of automatic EEG interpretation.
