53 Are Boys (names) Really Just Like Animals? Comparing Multi-Category Fluency Trials from the D-KEFS in Predicting Temporal Cortical Thickness in an Outpatient Memory Disorders Population

Abstract

Objective:Semantic fluency measures comprise a differing number of trials depending on the test battery and/or normative data used. Using semantic fluency trials from the Delis Kaplan Executive Function System (D-KEFS; Animals and Boys’ names), we sought to examine whether: 1) there was incremental benefit of multiple trials in associations with aggregated temporal cortical thickness and 2) patterns of neuroanatomical associations with specific temporal lobe structures differed between Animals and Boys’ names trials.Participants and Methods:Archival records of adults who completed a neuropsychological evaluation which included the semantic fluency measures of interest and had undergone structural MRI were identified (n=243, Mage=72.35 years, SDage=6.74, Female=46.9%). Cortical thickness values were obtained using FreeSurfer and averaged across sub-regions, separately for the left and right temporal lobe, per recommendations from the FreeSurfer group. Multiple linear regression models were fit to examine separate and incremental contribution of both Animals and Boys’ names, on temporal lobe thickness, including age, sex, and education in the models. Zero order correlations with each of the temporal cortical thickness areas (inferior, middle, and superior temporal; banks of the superior temporal sulcus, fusiform, transverse temporal, entorhinal, temporal pole, and parahippocampal cortices) were also computed to identify more focal neuroanatomical correlates.Results:Animals and Boys’ names trials individually accounted for a significant proportion of variance when predicting temporal cortical thickness over and above demographics, but Animals was a considerably stronger predictor for left temporal cortical thickness (Left: Animals AR2 =.127*, Boys’ names AR2 = .067*; Right: Animals AR2 =.074*, Boys’ names AR2 = .065*). The variance accounted for by Boys’ names incrementally over Animals was not significant (AR2 = .004 for left and .015 for right hemispheres, respectively). Similarly, though the composite Category fluency index accounted for a significant proportion of the variance independently, it did not add incrementally over and above Animals alone when predicting cortical thickness in either hemisphere. When examining simple correlations with specific temporal cortices, Animals consistently had correlations of a greater magnitude than Boys’ names within the left hemisphere (Animals r>.3 for superior, middle, inferior, and fusiform gyri; Boys’ names r< .3 for all cortical thickness regions). Greater variability was noted for associations with right temporal thickness but Animals continued to show associations of a greater magnitude of associations than Boys’ names for several sub-regions. * denotes significance at p < .01.Conclusions:The additional Boys’ names trial does not confer significant benefit over Animals alone, when predicting cortical thickness in either temporal lobe. Additionally, overall category fluency provided little incremental utility over and above the Animals trial alone in predicting temporal thickness. Psychometrically, it is expected that composites derived from multiple trials are more robust. However, this study demonstrates that it is important to examine whether the administration of additional trials is truly beneficial, particularly in a climate where brevity of neuropsychological assessment is critically desired. Further, psychometric tests have historically been validated against other neuropsychological measures, but it is critical we also validate measures against neuroanatomical correlates.