Search for a Common Cause Factor amongst Cognitive, Speed and Biological Variables Using Narrow Age Cohorts

Abstract

Background: Cross-sectional studies have demonstrated that age-related effects on many cognitive and biological functioning variables are shared. A number of theories postulate that these associations arise from a common causal mechanism responsible for change across a wide range of functions. However, it has been claimed that the finding of shared variance among variables in these studies may be spurious because most studies have been cross-sectional and based on age-heterogeneous samples. Objective: To examine evidence for changes with age in variance shared by cognitive variables using a narrow-age cohort design. Methods: Three samples of adults with age ranges 20-24 years (n = 2,404), 40-44 years (n = 2,530) and 60-64 years (n = 2,510) were drawn from a population-based study. These groups were supplemented by data from an older sample of individuals (77 years and over) from a second population-based study (n = 374). Four models of the structure of a range of cognitive, speed and biological variables derived from previous cross-sectional research were fitted to data. These models were complete independence of variables within a cohort, a single common factor, multiple independent factors, and a second order factor model. Results: Model fit and the magnitude of loadings from multiple-group confirmatory factor analyses was compared between age groups. The fit for all four models was worse in the 60-64 year age group. Factor loadings on the common factor specified in these models varied between the groups, but the speed factor rather than the biological factor showed increased loadings as a function of age. Conclusions: This narrow age-cohort study did not support the existence of developmental changes in the structure of cognitive and biological variables across the lifespan. Shared variation between measures of cognition and biological function may be an artifact of the research designs that have generated these findings. This casts doubt of the existence of a common causal mechanism. However, it may be that such changes manifest only in later old age.