The impact of aging on cognitive functions. An 11 year follow-up study of four age cohorts.

The literature on longitudinal studies of aging and cognitive performance is reviewed. The review reveals some ambiguity which seems primarily to be due to selection bias in the study groups and insufficient statistical modelling and parameterization. The main purpose of this study was to analyze the impact of aging on cognitive performance of the general adult population. In addition, the following problems are considered: If an aging impact is found, is it then evenly distributed across different age cohorts and different cognitive functions? Is the effect of different subject-related and situational factors on the test parameters stable over time? Is it possible to eliminate the impact of the examiner in cognitive testing? Moreover, do certain background factors such as age (generation), sex, schooling, etc. or test performance in an earlier examination have predicting power for the participation at a later stage in the study? The analysis was made using a longitudinal design with four successive age cohorts born in 1952, 1942, 1932, and 1922. 1,026 subjects were examined in 1982-83 and 711 were re-examined in 1993-94. At both occasions the subjects were tested by means of the Cognitive Function Scanner method. Attrition was found in all four age cohorts but most profound in the oldest cohort. Attrition occurred most frequently among subjects with only primary schooling and low social status. Age (generation), Sex, Schooling, Occupational Activity Status, Smoking Status, Psychic Stress, and questionnaire positive Cardiovascular Disease in the 1982-83 study all had significant relation to non-participation in the follow-up, either because of death, refusal to participate or because the subjects did not turn up on the agreed date for examination. Also a few test parameters, primarily in the attention tests and concentration tests, were found to be significantly related to non-participation in the follow-up. The results of the follow-up study, in which each subject was his own reference, showed two different tendencies. One tendency indicated some decline in performance. It was found in parameters for non-verbal learning and memory, retention of verbal memory, visuomotor and visuospatial speed, concentration, and reaction time. The other tendency indicated some improvement in performance in relation to verbal learning and memory, visuomotor and visuospatial precision, and to visual perception. It was a general tendency that performance becomes more scattered with increasing age. The follow-up study showed that Generation, Sex, and Schooling are the most important subject-related factors to take into consideration in the evaluation of test results. The significant impact of the Generation factor indicates that cognitive performance depends not only on aging but is also modulated to a large extent by generation specific factors. In spite of intensive training and special focus on communication it was evident that the situational Examiner factor plays a significant role in all of the cognitive tests. The strong impact of the Examiner factor makes the use of test norms questionable. Norms should be used only as rough guidelines. In the individual psychological assessment, focus should be on the subject's problem solving strategies. This can be reached by supplementing the traditional psychometric scores with a range of response process observations obtained by recording the natural steps of the subject's response flow combined with latencies for each step, and analyzing the response chains by pattern analysis. The computer is considered the most appropriate tool for recording response flows.