Abstract
Two hundred and seventeen geriatric EEGs were quantitatively analyzed in relation to their intelligence test performances. The subjects were divided into three graded intellectual levels according to their test score in the three diagnostic categories classified with respect to hemiplegia. By means of a novel method using auto-regressive analysis and multivariate analysis, the mean spectrum patterns were numerically compared and classified in terms of general pattern similarities. As the results were in good accord with their intelligence test performances, it was suggested that there were proportional changes of EEG along with the intellectual deterioration. In that occasion, changes in frequency and dumping characters of alpha activity made the most significant contribution. These results suggest that an assessment of intellectual level might be possible through the EEG examination applied in this study and that a prediction of prognosis might also be expected.