Longitudinal trajectories of self‐reported subjective cognitive decline in relation to objective cognitive performance in a population‐based cohort

Abstract Background Subjective Cognitive Decline (SCD) may represent the onset of cognitive decline before impairment on standard cognitive tests occurs (Jessen et al., 2014). Previous cross‐sectional studies were shown to have limited ability to capture objective differences in cognitive performance between groups with and without SCD (Cacciamani et al., 2017). Understanding the interplay between changes in subjective memory ratings and objective cognitive function over time may have utility in identifying people at risk of dementia. Method 370 cognitively normal (CN) individuals with complete cognitive data ( Table 1 for characteristics) from Insight 46, a longitudinal neuroscience sub‐study of the MRC National Survey for Health and Development (the British 1946 Birth Cohort), underwent cognitive, clinical, and physical assessments, and neuroimaging (combined MRI/18F‐Florbetapir‐PET). SCD was measured using MyCog (Rami et al., 2014) a validated tool, from the SCD‐Questionnaire, where higher scores indicate greater subjective complaints. Linear regression models were used to test whether longitudinal cognitive change as measured using the Preclinical Alzheimer Cognitive Composite (PACC) and a visual short‐term memory binding task were influenced by baseline MyCog and whether there are parallel associations between rates of change in MyCog and longitudinal cognitive change. Covariates were sex, age at baseline visit, childhood cognitive ability, educational attainment, and adult socioeconomic position. Result Baseline MyCog scores did not predict rates of change in PACC or visual short‐term memory binding outcome measures ( Table 2 ). Rates of MyCog change alone did not predict rate of change in PACC, localisation error, total correct number of trials or proportion of misbinding errors ( Table 3 ). However, we observed an interaction effect with baseline amyloid‐b status, whereby greater rates of subjective memory concerns predicted faster rates of decline in the proportion of misbinding errors in amyloid‐b positive participants only ( Table 3; Figure 1 ). Conclusion These findings suggest that higher amyloid‐b burden in CN individuals with subjective memory complaints is associated with faster rates of decline in visual working memory. Corroborating previous reports that feature binding tasks can objectify subtle cognitive deficits found in SCD (Koppara et al., 2015) and serve as a promising tool for identifying people at risk of dementia.