Pubertal development underlies optimization of inhibitory control through specialization of ventrolateral prefrontal cortex

Abstract

Inhibitory control improves into young adulthood after specialization of relevant brain systems during adolescence. However, the biological mechanisms supporting this unique transition are not well understood. Given that adolescence is defined by puberty, we examined relative contributions of chronological age and pubertal maturation to inhibitory control development. 105 8-19-year-olds completed 1-5 longitudinal visits (227 visits total) in which pubertal development was assessed via self-reported Tanner stage and inhibitory control was assessed with an in-scanner antisaccade task. As expected, percentage and latency of correct antisaccade responses improved with age and pubertal stage. When controlling for pubertal stage, chronological age was distinctly associated with correct response rate. In contrast, pubertal stage was uniquely associated with antisaccade latency even when controlling for age. Chronological age was associated with fMRI task activation in several regions including the right dorsolateral prefrontal cortex, while puberty was associated with right ventrolateral prefrontal cortex (VLPFC) activation. Furthermore, task-related connectivity between VLPFC and cingulate was associated with both pubertal stage and response latency. These results suggest that while age-related developmental processes may support maturation of brain systems underlying the ability to inhibit a response, puberty may play a larger role in the effectiveness of generating cognitive control responses.

Keywords: Adolescence; Antisaccade; Inhibitory control; PPI; Puberty; fMRI.

Fig. 1

Distribution of included participants across age (top) and pubertal stage (bottom), colored by sex.

Fig. 2

Diagram of experimental task paradigm, depicting structure of experimental run (A) and trial type (B). Originally from Velanova et al. (2008) and Ordaz et al. (2013).

Fig. 3

Plots of correct response rate data across age (years) using both a linear mixed model (A) and generalized additive mixed model (B). Both show a significant positive association, such that increasing age is associated with improvements in correct response rate. Bottom panel in (B) depicts intervals of significant age-related change. Individual datapoints reflect values for each session, and connected lines reflect sessions from the same participant.

Fig. 4

Plots of response latency data across mean pubertal stage using both a linear mixed model (A) and generalized additive mixed model (B). Both show a significant negative association, such that increasing pubertal stage is associated with decreasing response latency. Bottom panel (B) depicts intervals of significant age-related change. Individual datapoints reflect values for each session, and connected lines reflect sessions from the same participant. In (A), red and blue lines reflect females and males, respectively.

Fig. 5

Plots of inhibitory control regions-of-interest (ROIs) in which significant age or puberty effects were observed. Both age and puberty were tested in each ROI, but the measure producing the best fitting model is visualized. Age was the most robust predictor of change in the bilateral parietal eye fields (PEF), right dorsolateral prefrontal cortex (DLPFC) and dorsal anterior cingulate cortex (dACC), while puberty was most associated with the right ventrolateral prefrontal cortex (VLPFC). Individual data points reflect values for each session, and connected lines reflect sessions from the same participant.

Fig. 6

Maps showing R VLPFC PPI connectivity with the rest of the brain during correct trials. The top panel (A) shows unthresholded maps and the bottom (B) shows a cluster-corrected threshold of p < 0.05.

Fig. 7

Maps showing clusters in R VLPFC PPI connectivity analysis that were significantly associated with pubertal stage when controlling for sex (p < 0.05 cluster-corrected).

Fig. 8

PPI analyses revealed several clusters within the brain in which task-related connectivity with the right VLPFC was significantly associated with pubertal stage, including the cingulate cluster visualized above (p < 0.05 cluster-corrected). Connectivity beta values were z-scored before plotting their association with mean pubertal stage (A). Within this cluster, connectivity was significantly associated with response latency, such that greater VLPFC connectivity was associated with faster response latencies (β = −0.19, p_FDR = 0.04). There was also a significant effect of sex such that females had slightly slower response latencies across this sample’s age range (β = - 0.31, p = 0.03).

Fig. 9

Maps showing R DLPFC PPI connectivity with the rest of the brain during correct trials. The top panel (A) shows unthresholded maps and the bottom (B) shows a cluster-corrected threshold of p < 0.05 (bottom).

Fig. 10

Maps showing clusters in R DLPFC PPI connectivity analysis that were significantly associated with chronological age when controlling for sex at a cluster-corrected threshold of p < 0.05.