The Aberdeen birth cohorts of 1921 and 1936 (ABC21 and ABC36) were subjected to IQ tests in 1932 or 1947 when they were aged about 11y. They were recruited between 1997–2001 among cognitively healthy community residents and comprehensively phenotyped in a long-term study of brain aging and health up to 2017. Here, we report associations between baseline cognitive test scores and long-term cognitive outcomes. On recruitment, significant sex differences within and between the ABC21 and ABC36 cohorts supported advantages in verbal ability and learning among the ABC36 women that were not significant in ABC21. Comorbid physical disorders were self-reported in both ABC21 and ABC36 but did not contribute to differences in terms of performance in cognitive tests. When used alone without other criteria, cognitive tests scores which fell below the −1.5 SD criterion for tests of progressive matrices, namely verbal learning, digit symbol and block design, did not support the concept that Mild Cognitive Impairment (MCI) is a stable class of acquired loss of function with significant links to the later emergence of a clinical dementia syndrome. This is consistent with many previous reports. Furthermore, because childhood IQ-type data were available, we showed that a lower cognitive performance at about 64 or 78 y than that predicted by IQ at 11 ± 0.5 y did not improve the prediction of progress to MCI or greater cognitive loss. We used binary logistic regression to explore how MCI might contribute to the prediction of later progress to a clinical dementia syndrome. In a fully adjusted model using ABC21 data, we found that non-amnestic MCI, along with factors such as female sex and depressive symptoms, contributed to the prediction of later dementia. A comparable model using ABC36 data did not do so. We propose that (1) MCI criteria restricted to cognitive test scores do not improve the temporal stability of MCI classifications; (2) pathways towards dementia may differ according to age at dementia onset and (3) the concept of MCI may require measures (not captured here) that underly self-reported subjective age-related cognitive decline.
Bibliographical noteAcknowledgments: We remain grateful to the kindness of the staff at the Scottish Council for Research in Education who allowed us access to their archive and remained supportive and gracious throughout our collaboration. We thank the many people of Aberdeen who volunteered generously and committed to the long-term success of this program. We thank Victoria Bourne, who made substantial contributions to study design, data collection, data analysis and hypothesis generation. Jen Herbert (deceased) recruited the ABC36 participants, collected data (sessions I and II) and, through her encouragement and professionalism, ensured the continued involvement of many participants. She was much loved by participants and the study team.
Funding: The Aberdeen Birth Cohort 1921 and 1936 research program was established in 1997 with funding from the Henry Smith (Kensington Estates) Charity and continued by The UK Biotechnology and Biological Sciences Research Council (1999–2002), The Wellcome Trust (2001–2006), The Scottish Government (2000–2002), the Medical Research Council (2003), Alzheimer Research UK (2002–2005) and the University of Aberdeen Development Trust (2007–2010, 2014).
- Mild Cognitive Impairment (MCI)
- cognitive test
- cognitive reserve
- childhood IQ
- longitudinal study
- normative data