In 1987, neuropathologists representing the charter group of CERAD clinics met to develop a practical instrument for standardizing the postmortem evaluation of the brains of individuals with Alzheimer's disease (AD) who had been enrolled in the multicenter CERAD study. At that time, standardized CERAD assessment batteries of the clinical and neuropsychological characteristics of this disorder had been established, and the need for a comparable evaluation of the neuropathologic findings had become apparent. Two years earlier, a panel of neuropathologists had recommended a set of minimal microscopic criteria for the diagnosis of AD.I They suggested that the diagnosis be based on an agerelated minimum number of senile plaques per microscopic field of neocortex. Although that report provided useful guidelines that had not been previously available, the criteria were arguably arbitrary and did not take into account differences in the stains and techniques used and variations in interpretation.For these reasons, the CERAD neuropathologists formed a task force to develop more accurate and reliable neuropathologic criteria for AD, determine the neuropathologic spectrum of this disorder, and establish the types and frequency of coexisting disorders. A standardized neuropathologic protocol that involved a semiquantitative approach to assessment of the frequency of senile plaques, neurofibrillary tangles, and other changes was formulated and tested.2 The frequency of neocortical neuritic plaques is correlated with the patient's age to arrive at an "agerelated plaque score," which, along with the clinical history of dementia, is used to establish a level of certainty for the diagnosis of AD (e.g., definite AD, probable AD, possible AD, and normal brain). The neuropathologic protocol includes a data entry form and illustrated guidebook with photomicrographs.To further promote standardization, 24 neuropathologists from 18 medical centers participated in a study to determine the inter-rater reliability among laboratories for the neuropathologic evaluation of AD.3 Autopsy brain tissue from eight patients with a clinical diagnosis of AD and two control subjects was assessed. Contiguous unstained paraffin sections from the frontal cortex of these 10 cases were distributed to neuropathologists for evaluation of AD histopathologic changes, using the stains of their choice. Rank order determinations of the relative severity of AD changes in the 10 cases revealed an inter-rater agreement of 75%. Significant differences were noted among raters for quantitative plaque and tangle counts, reflecting varying stain sensitivity, technique, and interpretation. Reasonable inter-rater agreement was found, however, using the semiquantitative measures of CERAD. Each neuropathologist received feedback on his or her data compared with those of the group. This exercise heightened the sensitivity of neuropathologists to the importance of using sensitive stains and uniform measures. Moreover, the results of this study bolstered the decision of the CER...