Subjective Cognitive Decline Is More Accurate When Metamemory Is Better

Subjective cognitive decline (SCD) has emerged as one of the first manifestations of Alzheimer's disease (AD). However, discrepancies in its relationship with tests of memory and other cognitive abilities have hindered SCD's diagnostic utility. Inter-individual heterogeneity in metamemory, or memory awareness, and the use of clinical measures of cognition lacking sensitivity to early cognitive dysfunction, may contribute to these discrepancies. We aimed to assess if the relationship between SCD and markers of early cognitive dysfunction is moderated by metamemory abilities.

The sample included 79 cognitively healthy older adults (77% female, 68% White, and 32% Black participants) with a mean age of 74.4 (SD = 6.1) and 15.9 (SD = 2.7) years of education. Metamemory was assessed using an episodic Feeling of Knowing test with four 5-item trials. Outcome measures included a resolution metric defined as a gamma correlation reflecting the accuracy of item-level predictions ("Will you know the correct answer?"). Early cognitive dysfunction was measured through the Loewenstein-Acevedo Scale for Semantic Interference and Learning (LASSI-L) and the Short-Term Memory Binding Test (STMB), measures sensitive to preclinical AD. SCD was assessed with a 20-item questionnaire that asked participants to compare themselves to others their age on a 7-point Likert scale. Regression analyses examined whether a potential relation between SCD and early cognitive dysfunction was moderated by metamemory.

Subjective cognitive decline was associated with susceptibility to semantic proactive interference such that greater complaints were associated with increased susceptibility to semantic proactive interference (b = -0.30, p = 0.003) only. Metamemory moderated the association between SCD and susceptibility to and recovery of semantic proactive interference such that those with more accurate metamemory showed a stronger association between increased complaints and susceptibility to semantic proactive interference (b = -0.71, p = 0.005; b = -0.62, p = 0.034). Metamemory, however, did not moderate the association of SCD with retroactive semantic interference nor short term memory binding.

The accuracy of an individual's metamemory, specifically their ability to adjust moment to moment predictions in line with their performance, can influence the extent to which SCD maps onto objective cognition. Such self-referential assessment should be considered when interpreting SCD.