Components of Executive Function Predict Regional Prefrontal Volumes

Abstract ObjectiveDesigned to measure a diversity of executive functioning (EF) through classical neuropsychological tests, the Delis-Kaplan Executive Function Scale (D-KEFS) allows for the investigation of the neural architecture of EF. We examined how the D-KEFS Tower, Verbal Fluency, Design Fluency, Color–Word Interference, and Trail Making Test tasks related to regional frontal lobe volumes, quantifying how components of EF were represented in disparate neural networks. MethodAdults from the Nathan Kline Institute – Rockland Sample (NKI-RS), an open-access community study of brain development, with complete MRI (3T scanner) and D-KEFS data were selected for analysis (N = 478; ages 20-85). In a mixed-effects model predicting volume, D-KEFS task, D-KEFS score, region of interest (ROI; 13 frontal, 1 occipital control), were entered as fixed effects with intercepts for participants as random effects. Results“Unitary” EF (average of D-KEFS scores) was positively associated with superior frontal, rostral middle frontal, and lateral orbitofrontal volumes; a negative association was observed with frontal pole volume (| z-score slope | range = 0.040 to 0.051). “Diverse” EF skills (individual D-KEFS task scores) were differentially associated with two or three ROIs, respectively, but to a stronger extent (| z-score slope | range = 0.053 to 0.103). ConclusionsThe neural correlates found for the D-KEFS support the prefrontal modularity of EF at both the unitary (aspects of EF ability common to all tasks) and task (diverse EF skills) levels. The separation of task-general variance in neurocognition from task-specific variance can further evaluate neuropsychological tests as indices of brain integrity. Public Significance StatementsOur results support the relationship between larger lateral prefrontal cortex and greater executive function.A composite of executive function performance, which was broadly associated with the prefrontal cortex, may be ideal for assessing diffuse frontal lobe damage (e.g., hypoxia).Individual executive functions, which were more narrowly but strongly related to specific prefrontal regions, could be better for assessing the effects of localized brain injuries (e.g., tumor).