Analysis of age-related cognitive functioning using reliability and time to Failure Rate Theory

Although cross-sectional (between-person) data for neuropsychological tests are widely used for the assessment of dementia, they have not been interpreted in the context of a cohesive developmental theory. Some researchers have even considered these data misleading because they do not provide a consistent age-cognition relationship. The present study was designed to reexamine published neuropsychological norms using the Weibull Cumulative Distribution Function, a curve fitting method often used in Reliability and Time to Failure theory. The authors demonstrate that the application of the Weibull Cumulative Distribution Function is useful for describing and characterizing the mean decay and relative resilience of different cognitive abilities as a function of age. It also illustrates the importance of using theoretically based nonlinear regression models rather than traditional linear or quadratic models for the understanding of neuropsychological test norms. Finally, it shows the importance of isolating specific cognitive domains within neuropsychological tests rather than using summary indexes. The results are discussed in the context of brain reserve models, notions regarding the modularity of the brain, cross-sequential studies of cognitive change, and structural and functional imaging studies.