Association between recollection and familiarity‐based memory measures and plasma biomarkers for Alzheimer’s disease

Abstract

AbstractBackgroundNormal aging and preclinical Alzheimer’s Disease (AD) share common features, such as structural changes to medial temporal lobe regions and declining episodic memory. However, prior work suggests qualitative differences between AD‐related memory changes and those due to aging. Specifically, recollection‐based, or contextual, memory declines with both normal aging and AD. Alternatively, familiarity‐based, or item, memory is more specific to preclinical AD and atrophy in the perirhinal cortex. In the current analysis, we examined the relationship between these forms of memory with plasma biomarkers associated with AD and neurodegeneration.MethodAt the Penn Alzheimer’s Disease Research Center, 127 cognitively‐normal participants (age=73.12±6.62, 59% female, years of education=16.40±2.62, 80% Caucasian) completed an experimental memory paradigm assessing recollection and familiarity along with standard neuropsychological testing. In this paradigm, pictures of common objects were presented in 80 pairs at study and participants were told to pick the object they preferred. At test, 40 intact, 40 re‐arranged, and 40 novel object pairs were presented. Recollection was defined as the probability(intact)–probability(rearranged), while familiarity was defined as the probability(rearranged)/(1‐R). To account for differences in base rates of false alarms (‘‘old’’ responses to novel pairs), familiarity was calculated using a measure of discrimination (d’) derived from signal detection theory. Whole blood samples were collected in EDTA‐tubes, spun‐down, and stored at ‐80°C. Measures of glial fibrillary acidic protein (GFAP), neurofilament light chain (Nfl), and plasma tau (p‐Tau181) were generated with the Simoa HD‐X analyzer system.ResultFamiliarity correlated with GFAP (r(125) =‐0.24, p<0.05) (Figure 1) and NfL (r(125) =‐0.23, p<0.05) (Figure 2). Further, a partial correlation controlling for age and days between the experimental and plasma measure remained significant for GFAP (r(125)=‐0.25, p<0.05). Additionally, recollection correlated with Nfl (r(125) =‐0.25, p<0.05) (Figure 3). Correlations with standard neuropsychological tests revealed no significant results.ConclusionThis study reveals that within the normal aging spectrum, familiarity‐based memory correlates with molecular measures that have been associated with amyloid (GFAP), while both measures correlated with a non‐specific measure of neurodegeneration (NfL). This supports the notion that familiarity may be sensitive to preclinical AD.