Relationship between cortical thickness and fluency in the memory disorders clinic population

Abstract

As brain networks break down in neurodegenerative disorders such as Alzheimer's disease, language and executive function frequently decline. We aimed to quantify the relationship between fluency (both categorical and phonemic) and cortical thickness using data from a large cohort of patients who were undergoing assessment at a memory disorders clinic. In addition, we defined the pattern of these relationships across the four major lobes of the brain.A total of 590 patients underwent extensive psychometric testing, including categorical (animal-naming) and phonemic (FAS) tests of fluency. All patients also underwent structural MRI featuring a volumetric T1-weighted sequence that served as the input for postprocessing calculations using FreeSurfer, yielding cortical parcellations and thicknesses. The fluency-thickness relationships were summarized using Pearson's correlation coefficient.In a univariable analysis over all lobes, there were significant correlations using categorical fluency with both cortical thickness and age, with education less correlated; using phonemic fluency there were similar correlations with cortical thickness and age, but education was more correlated. Neither handedness nor sex was significantly correlated with either categorical or phonemic score.At a lobar level, for both fluency tests, scores were positively correlated with cortical thickness in all lobes; these relationships were strongest in the temporal lobe (p < 0.01). The correlations for categorical testing were generally stronger than the correlations for phonemic testing and were again strongest in the temporal lobe (r = 0.38 for categorical testing vs 0.22 for phonemic testing). The bilateral parietal lobes were more important for categorical testing than for phonemic testing, and the left frontal lobe was more important for phonemic testing than for categorical testing. Comparison of the homologous lobes between the two hemispheres demonstrated that only the frontal lobes were significantly different for both scores, with the left side having a stronger relationship with the scores (categorical: r = 0.21 for left; r = 0.14 for right; p < 0.01. phonemic: r = 0.13 for left; r = 0.08 for right; p < 0.01).In conclusion, these results demonstrated that structural MRI and fluency tests reveal a significant spatial pattern of correlations between cortical thickness and fluency, which varies with the type of fluency test.