Background:
•Accelerated Long-term Forgetting (ALF) is the phenomenon whereby material is retained normally over short intervals (minutes or hours) but forgotten abnormally rapidly over longer periods (days or weeks).
•There is evidence that ALF is an early marker of cognitive decline in preclinical Alzheimer’s disease (AD) [1], but questions remain about its relationships with biomarkers of preclinical Alzheimer’s disease pathology.
•We previously reported associations between some measures of ALF and amyloid-positivity and brain atrophy rates in cognitively normal 73-year-olds in the Insight 46 study – a sub-study of participants from the MRC National Survey of Health and Development (also known as the British 1946 Birth Cohort) [2]. Participants were all born during the same week and have completed regular assessments of their health and cognition throughout childhood and adulthood.
•We now report investigation of the relationships between AD biomarkers and ALF from the next time-point of this study at age ~78 years.
Methods:
•At age ~78, participants (n=613) completed the ALF assessment comprising verbal and visual memory tasks: the 12-item Face-Name Associative Memory Exam (FNAME-12) [3,4] and Complex Figure Drawing [2] (Fig.1, Fig.2) ALF scores were calculated as % of material retained after 7 days, relative to 30 minutes.
•Plasma p-tau217 was quantified ~8 years earlier (at age ~70 years) and categorised as low, intermediate or high using published cut-points [5]. A sub-sample of participants (n=307) had amyloid-PET status determined at age ~73, and rates of whole-brain atrophy (derived from serial MRI) between ages ~70-78.
•Linear regression models were used to investigate associations between memory performance (trial-by-trial and ALF) and, separately, amyloid status (positive, negative), p-tau217 (low, intermediate, high) and brain atrophy rates, adjusting for relevant covariates. Analyses were repeated including only participants who were cognitively normal at age ~78 (Clinical Dementia Rating (CDR) =0), to investigate whether subtle ALF was detectable in this group.
Results:
•Participants (n=613) were 49.9% female with mean [SD] age of 77.6 [1.2] years. Prevalence of low, intermediate and high p-tau217 were 75.4%, 17.1% and 7.5%. The intermediate and high-p-tau217 groups both performed worse on the verbal and visual memory tests, relative to the low-p-tau217 group. These differences were seen even on the immediate and 30-minute recall tests (Fig.3A, 3B).
•In the sub-sample with prior neuroimaging, amyloid-positivity (prevalence 30.5%) was associated with poorer verbal and visual recall ~4 years later (Fig.3C, 3D). Faster rates of brain atrophy were associated with the same outcomes (Table 1). Again, these effects were seen even on the immediate and 30-minute recall tests.
•Although coefficients for ALF scores were in the expected direction, they were only statistically significant on FNAME-12 for faster brain atrophy.
•When including only participants with CDR=0 (prevalence 81.2%), associations with memory were attenuated for p-tau217 and amyloid-positivity; however, associations with brain atrophy rates remained, including for ALF score on FNAME-12 (data not shown).
Conclusions:
•Our previous work [2] found evidence of subtle ALF at age ~73 in individuals with biomarkers of preclinical AD. However, by age ~78 years, memory deficits were apparent in immediate and short-delay recall in individuals with biomarkers of Alzheimer’s pathology measured up to ~8 years beforehand.
•Among those who were cognitively normal (CDR=0) at age ~78, faster rate of brain atrophy was associated with ALF on a challenging associative memory test.
•These results support the utility of ALF tests for detecting early memory impairment in preclinical AD, but suggest that the prime window for this in population-based samples may be before individuals reach their late 70s.
References: 1. Pizzoni et al. (2026) JAD | 2. Lu et al. (2025) Brain | 3. Papp et al. (2014) Clin. Neuropsychol. | 4. Lu et al. (2026) Neuropsychology | 5. Ashton et al. (2024) JAMA Neurol.