Discriminant Analysis of Electroencephalograms Recorded from Renal Patients

Cassification of electroencephalograms (EEGs) recorded from renal patients and normal subjects by discriminant-analysis-based techniques is described. The method for generation of two discriminant functions and a classification space is discussed and the classifier is tested on a randomly selected set of normal subjects and azotemic patients. The results of the discriminant function classification method are compared with another previously used method for assessing the amount of slow-wave-related EEG activity that is correlated with increasing renal failure. Finally, the discriminant functions are used to track a single patient from azotemia to dialysis and transplant.