Relative influence of subject variables and neurological parameters on neuropsychological performance of adult seizure patients.

We present the results of a study examining relationships between subject characteristics (age, education, sex, race) and neurological factors on neuropsychological test performance of adult seizure patients. Test scores from a modified and expanded Halstead-Reitan test battery for 250 epileptics served as criterion variables for sets of multiple regression analyses using subject characteristics and neurological parameters as predictors. Initially, subjects were divided into two subsamples and separate analyses for subject and neurological variables were pursued. Double cross-validated results indicated that test performance was significantly related to both sets of predictor variables, yet the magnitudes of relationship were rather consistently highest for demographic attributes. The relative contribution of individual predictor variables subsequently was explored in a second set of multiple regression analyses examining both subject and neurological variables within the combined patient sample. Results indicated that one or more subject characteristic plus one or more neurological variable accounted for a significant amount of test variance for almost all of the measures under consideration. A subject variable was entered first in step-wise regression analyses for 78 % of the measures, with education proving to be the most potent predictor of test status. In general, these data support the need for test norms which make adjustments for nonneurological subject attributes.