Free recall versus combined cued recall and recognition to detect true memory impairment on the Montreal Cognitive Assessment

Abstract

AbstractBackgroundThe Montreal Cognitive Assessment (MoCA) is widely used as a brief screening measure to characterize overall cognitive status. Although only delayed free recall of the brief 5‐item memory list contributes to widely used MoCA total scores, the optional cued recall and multiple‐choice recognition subtests may provide better diagnostic accuracy than free recall alone.MethodData on 719 individuals with Mild Cognitive Impairment and 601 cognitively unimpaired controls were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. The Rey Auditory Verbal Learning Test (AVLT) delayed free recall condition was used to characterize level of memory performance. Participants with demographically adjusted T‐scores of ≤ 2 SDs below the mean were classified as ‘impaired.’ Binary logistic regressions assessed if combined MoCA cued/recognition performance was a significant predictor of impaired delayed recall on the AVLT beyond the contribution of free recall and while covarying for age, education, and gender. Sensitivity, specificity, and likelihood ratio values were also examined for the MoCA free recall and combined cued/recognition scores.ResultFree recall on the MoCA was a significant predictor of delayed recall on the AVLT (b = ‐1.207, Wald χ2 (1) = 128.044, p < 0.001, OR = 0.299, 95%CI = [0.243, 0.369]. The addition of the combined MoCA cued recall/recognition conditions improved the overall model fit (χ2(1) = 30.616, p<.001) and predicted impaired AVLT recall such that for every additional word on combined MoCA cued recall/recognition, the likelihood of being in the AVLT memory impaired group decreased by 33% (b = ‐0.400, Wald χ2 (1) = 29.834, p < 0.001, OR = 0.670, 95%CI = [0. 581, 0.774]). Combined MoCA cued recall/recognition had both higher specificity and likelihood ratios in detecting AVLT memory impairment than MoCA free recall, while higher sensitivity values were present for MoCA free recall (see Table 1).ConclusionUse of combined cued/recognition memory conditions provides clinicians with improved diagnostic accuracy in detecting memory impairment. In settings where specificity in detecting memory impairment is important, administration of the MoCA free recall as well as cued recall and recognition conditions is recommended.