Aspects of Declarative Memory Functioning in Adulthood Cross-Sectional and Longitudinal Studies

The general objective of the thesis was to examine aspects of declarative memory functioning across the adult life span. The four papers were based on data collected as part of the Betula Prospective Cohort Study (Nilsson et al., 1997) and included largescale population-based samples of participants in the age range 35 to 90. In study I and study II the possibility that age differences in episodic memory may be compensated for by provision of encoding support in the form of enactment was investigated, using free and cued recall and recognition portioned into components of recollective experience as the dependent measures. In Study III, unitary, two-, and multi-factorial models of declarative memory were compared and age-invariance was tested for. In Study IV cross-sectional age differences were contrasted with five-year longitudinal changes on aggregate measures of episodic and semantic memory within age groups ranging from 35 to 85 years. The results of Study I and Study II demonstrated that enactment constitutes an effective form of encoding support, but that the age differences generalize across this form of encoding support. Study II indicated that most of the age-related variance in recognition and levels of recollective experience following enacted and non-enacted encoding was shared by a measure of processing speed. Study III confirmed that a two-factor model of declarative memory (episodic and semantic memory) yields superior fit as compared with a unitary model of declarative memory. However, the best fitting model was a six-factor model with recall and recognition (episodic memory) and knowledge and fluency (semantic associated with different patterns of age-related differences, with some indications that the first-order factors show differential age-related patterns, indicative of variability that cross-sectional data may give a false impression of decline for adults in the age range 35-60 years for episodic memory. There was no evidence of time-related decline within these age groups, even though practice effects were taken into account. However, past this age, substantial time-related decline was observed for the older adults, in line with cross-sectional data. Semantic memory performance tended to improve across time for the younger groups, but decline in old age, although the magnitude of this decline was less pronounced than for episodic memory. Cohort differences in education may be one important factor underlying the discrepancy between t