This thesis develops a model of predictors of cogilltive aging based on an analysis of the empirical literature on predictors and the theory that cognitive aging is concommitant with biological aging. This model and its variants are then tested empirically in a cross-sectional study of older women. Inferences are then drawn from the empirical study to support the author's multiple-cause-central-factor theory of cognitive aging.
In particular, the study reported here sought (a) to demonstrate the utility of the construct of biological age for models of predictors of cognitive aging, and (b) to extend and integrate previous research by including cognitive, contextual and sensorimotor and physiological predictors in the one model.
The sensorimotor and physiological variables were conceptualised as biological markers of aging (biomarkers) and were used as indicators of a latent variable of biological age in a structural equation model of predictors of cognitive aging. The other non-cognitive predictors were contextual variables (health, activity and education) that have been investigated in previous research. The cognitive predictors included perceptual speed and working memory. These two cognitive predictors have been shown to explain agerelated variance in fluid abilities in previous research (Salthouse, 1992a).
On the basis of the literature review it was hypothesised that the biomarkers would explain age-related variance in cognition and show strong relationships with fluid abilities. Working memory and perceptual speed were also expected to explain age-related variance in fluid abilities. In comparison, the literature review suggested that contextual variables would explain general individual differences in cognition but only a very small proportion of agerelated variance in cognition.
One-hundred and eighty community dwelling, female volunteers aged 60 to 90 were assessed on a 3 hour battery of cognitive, sensorimotor and physiological measures and completed questionnaires on health and lifestyle. A young comparison group was also assessed on the same battery and questionnaires.
Structural equation analyses confirmed the hypothesis that the biomarkers formed a latent variable (Bioage) which mediated the entire effect of age on cognition. Health had an indirect effect on cognition via Bioage. Education and mental activity were important predictors of crystallised intelligence and age-independent variance in fluid abilities. Physical activity and mental health did not emerge as predictors of cognition in this study. Although raw correlations between the biomarkers and cognition were larger in the old-old than in the young-old, the structural paths among Bioage, Age, Physical Health, and cognition did not differ between age groups in a multigroup analysis. Working memory and perceptual speed both mediated age-cognition relations. Perceptual speed also mediated the relationship between the biomarkers and cognition showing that the biomarkers and perceptual speed shared age-related variance in cognition. Analysis of a young comparison group showed that the relationships between the variables used as biomarkers in the old group and cognition were much smaller but not absent altogether.
The findings of this study support the view that there is a central aging factor that accounts for age differences in cognitive performance in the latter part of the life span. The evidence for the central factor is the finding that a range of predictors share age-related variance in cognition. Some predictors - such as vision, hearing and perceptual speed are direct indicators of the aging brain but others, such as vibration sense and forced expiratory volume are not direct measures of the aging brain. Therefore the author adopts the view that there are several causes of the central factor involved in cognitive aging and proposes a multiple-cause-central-factor theory.