Three age groups (20-31, 37-49, and 55-67 years) of healthy men and women were compared on tests from the Halstead-Reitan neuropsychological test battery. Age-cohort differences in performance were observed for every measure (except Finger Tapping) and for overall indices of cognitive functioning, the Average Impairment Rating (AIR) and the Halstead-Reitan Impairment Index (HR II). When traditional cut scores (developed with age-heterogeneous samples) were used, alarmingly high percentages of subjects in the oldest group were defined as brain damaged. These findings were not attributable to higher levels of psychometrically-assessed anxiety and depression in the older groups or for the impaired subjects. Slightly more conservative criteria (fewer diagnoses of brain impairment) were then employed. The first, AIR greater than or equal to 2.00, resulted in a low and more reasonable base rate of brain impairment for the oldest group. The second, AIR greater than 2.00, resulted in age-parity in incidence of brain impairment. The heuristic value of more conservative AIR scores was discussed as were possible uses in applied clinical research settings. The limitations of this approach were noted with regard to its exclusive focus on the problem of false positive diagnoses. It was emphasized that, from a neuropsychological diagnostic perspective, cut scores based on healthy aging samples must ultimately be tested in the context of traditional neuropsychological studies in which external validation with neurosurgical and neurological criteria is employed.