Brain Network Correlates of Adolescent Interference Control

• ¹ The University of Melbourne, Melbourne Neuropsychiatry Centre, Australia
• ² Monash University, Monash Biomedical Imaging, Australia
• ³ The University of Melbourne, Melbourne School of Psychological Sciences, Australia

Background

Understanding the neural basis of teenage self-control is a central goal of developmental neuroscience. Previous work has emphasised prefrontal cortex regions (PFC), but recent functional magnetic resonance imaging (fMRI) evidence suggests three alternatives: 1) activation and functional connectivity of the cognitive control network (CCN); 2) deactivation and functional connectivity of the default-mode network (DMN); or 3) a dynamic interplay between the two systems. We investigated these hypotheses by studying how individual differences in interference control were related to task-related activation, deactivation, and functional connectivity. We also determined if similar relationships could be found in the resting-state.

Methods

Seventy-three 16-year-old adolescents were selected from a longitudinal study of development. Each subject performed the Multi-Source Interference Task (MSIT) and underwent a resting-state acquisition. A novel graph theoretic approach was used to correlate the MSIT reaction-time interference effect (IE) with task-related brain activation and deactivation, task-related functional connectivity, and resting-state functional connectivity. Behavioural relationships with higher-order brain network organisation were then investigated with modularity analyses.

Results

Significant IE correlations with fMRI task activity were found in the dorsal anterior cingulate (dACC), frontal eye fields (FEF), the intra-parietal sulci (IPS), lateral occipital cortices (LO), and two DMN regions. Task-related functional connectivity analysis demonstrated that these areas were embedded within an IE subnetwork containing 35 CCN and DMN regions centred on the superior colliculus (SC). Performance was also associated with the successful integration of IE subnetwork regions into two core modules during the transition from rest to task conditions. No IE relationships were found during the resting-state.

Discussion

Adolescent interference control was mediated by task-dependent recruitment of a large-scale subnetwork that included CCN and DMN regions. The subnetwork contained canonical PFC control areas (e.g., the dACC) in addition to visual attention (e.g., FEF, IPS, SC, and LO) and DMN regions. Together, the results emphasise the behavioural importance of task-dependent activation, deactivation, connectivity, and modular changes. This conclusion highlights the need to look beyond the frontal lobes to better understand teenage self-control.