Visual and spectral EEG analysis in the evaluation of the outcome in patients with ischemic brain infarction.

Serial EEGs were recorded in 15 patients with acute cerebral infarctions in order to study clinical and prognostic correlations. The EEG was recorded within 48 h from the first symptoms and thereafter weekly for 4 weeks. The EEGs were analyzed both visually and with a computerized spectral analysis. Eight of the patients recovered fully and seven had permanent neurological deficits. On admission, 87% of the patients had an abnormal EEG by visual analysis. The spectral parameters correlated well with visual findings, especially the delta and alpha bands. The spectral analysis was superior to visual in predicting the correct laterality of the lesion. It showed the correct side of the lesion in 87%, while the visual did it in only 54% of the cases. The first EEG records reliably predicted the outcome of the patients. The degree of background abnormality was most important in visual EEG analysis. In spectral analysis, parameters from single derivations were superior to the average of all derivations. A high proportion of delta or low proportion of alpha power were reliable indicators of poor outcome.