Specific memory binding indexes and their potential to discriminate between Alzheimer´s disease biomarker‐positive and biomarker‐negative older adults with amnestic mild cognitive impairment

Abstract

AbstractBackgroundThe Memory Binding Test (MBT) and Face‐Name Associative Memory Exam (FNAME‐12) are challenging tests based on relational memory binding paradigm. Recent studies support their sensitivity to subtle memory changes; however, their specificity to Alzheimer´s disease (AD) related memory deficit has not been studied. This study aimed to explore differences in MBT and FNAME‐12 scores between patients with amnestic mild cognitive impairment (aMCI) due to AD (aMCI‐AD), those with aMCI with non‐Alzheimer´s etiology (aMCI‐nonAD), and cognitively normal older adults (CN) as a control group, and to compare the potential of the challenging tests and the standard Rey Auditory Verbal Learning Test (RAVLT) to discriminate between the aMCI groups.MethodSixty‐seven Czech Brain Aging Study participants (aMCI, n = 37; CN, n = 30) underwent neuropsychological assessment (including the MBT and FNAME‐12), laboratory evaluation, and brain magnetic resonance imaging, and were further classified as aMCI‐AD (n = 18) or aMCI‐nonAD (n = 19) based on the results of amyloid positron emission tomography or AD cerebrospinal fluid biomarkers. The groups were compared using the multivariate analysis of covariance controlling for sex, age, education, and global cognition.ResultIn MBT, aMCI‐AD patients scored the lowest on the paired cued recall scores (CR) (aMCI‐AD<aMCI‐nonAD<CN, ps≤0.003). In FNAME‐12, aMCI‐AD patients scored the lowest on the CR of face‐occupation pairs (aMCI‐AD<aMCI‐nonAD<CN, ps<0.001). On the CR of face‐name pairs, both aMCI groups scored lower than CN (p<0.001); however, they did not differ from each other [p = 0.525; (aMCI‐AD = aMCI‐nonAD)<CN]. In RAVLT, aMCI‐AD patients did not differ from the aMCI‐nonAD patients, except for the delayed free recall (FR) (p<0.001). Adding the RAVLT delayed FR as covariate, the differences in the MBT and FNAME‐12 paired CR scores remained significant. To discriminate aMCI‐AD from aMCI‐nonAD, the MBT paired CR and FNAME‐12 CR of face‐occupation pairs yielded areas under the curve (AUCs) of 88% and 85%, respectively, which was comparable to RAVLT delayed FR (87%). Controlling for RAVLT, the AUCs were 91% for MBT and 88% for FNAME‐12.ConclusionOur findings support excellent discriminative ability of selected MBT and FNAME‐12 scores for AD‐related memory deficit and suggest their potential for differential diagnosis of memory deficit in older adults in addition to the standard RAVLT.