Umeå universitets logga

The objective of the study was to examine the degree of stability of interindividual differences in general cognitive ability (g) across the adult life span. To this end, we examined a sample of men (n = 262), cognitively assessed for the first time at age 18 (conscript data). The sample was reassessed at age 50 and at five year intervals up to age 65. Scores from conscript tests at age 18 were retrieved and three of the subtests were used as indicators of g in early adulthood. At age 50–65 years, four indicators served the same purpose. At the 15-year follow-up (age 65) two working memory measures were administered which allowed examination of the relationship with working memory capacity. Results from structural Equation Modelling (SEM) indicated extremely high level of stability from young adulthood to age 50 (standardized regression coefficient = − 95) as well as from age 50 to age 55, 60 and 65 with stability coefficients of .90 or higher for the for the latent g factor. Standardized regression coefficients between young-adult g and the g factor in midlife/old age were .95 from age 18 up to age 50 and 55, .94 up to age 60, and .86 up to age 65. Hence, g at age 18 accounted for 90–74% of the variance in g 32–47 years later. A close association between g and working memory capacity was observed (concurrent association: r = .88, time lagged association: r = .61). Taken together, the present study demonstrates that interindividual differences in g are extremely stable over the period from 18 to midlife, with a significant deviation from unity only at age 65. In light of the parieto-frontal integration theory (P-FIT) of intelligence, consistent with the close association between g and working memory capacity, midlife may be characterized by neural stability, with decline and decreased interindividual stability, related to loss of parieto-frontal integrity, past age 60.