Abstract
Animal fluency is a commonly used neuropsychological measure that is used in the diagnosis of amnestic mild cognitive impairment (aMCI) and Alzheimer disease. Although most individuals with aMCI have clinically normal scores on this test, several studies have shown that aMCI individuals' performance is significantly lower than that of cognitively unimpaired (CU) individuals. The aim of this meta-analysis was to characterize the effect size of animal fluency performance differences between aMCI and CU individuals. Literature search with search terms used were: "animal fluency and mild cognitive impairment," "semantic fluency and mild cognitive impairment," "category fluency and mild cognitive impairment." Both the standardized mean difference and the raw mean difference were derived from random effects analyses. Demographically adjusted z-scores for animal fluency performance for the aMCI groups were obtained to determine normative performance. Nineteen studies were included in the analysis. The standardized mean difference for animal fluency performance between CU and aMCI was 0.89 (95% confidence interval: [0.73; 1.04], P <0.001), I2 =70.3% [52.7%; 81.4%], which reflects a large effect size with moderate heterogeneity. The raw mean difference was -4.08 [-4.75; -3.38], P <0.001. The mean animal fluency z-score for aMCI groups was in the Low Average range (z=-0.77). This study found a substantial difference in animal fluency performance between aMCI and CU individuals. The aMCI groups' normative performance did not fall into the impaired range, indicating that there are important subclinical differences in animal fluency performance that may inform the design of cognitive end points for Alzheimer's disease prevention trials.