Neurodevelopmental shifts in learned value transfer on cognitive control during adolescence

Abstract

Value-associated cues in the environment often enhance subsequent goal-directed behaviors in adults, a phenomenon supported by the integration of motivational and cognitive neural systems. Given that the interactions among these systems change throughout adolescence, we tested when the beneficial effects of value associations on subsequent cognitive control performance emerge during adolescence. Participants (N = 81) aged 13–20 completed a reinforcement learning task with four cue-incentive pairings that could yield high gain, low gain, high loss, or low loss outcomes. Next, participants completed a Go/NoGo task during fMRI where the NoGo targets comprised the previously learned cues, which tested how prior value associations influence cognitive control performance. Improved accuracy for previously learned high gain relative to low gain cues emerged with age. Older adolescents exhibited enhanced recruitment of the dorsal striatum and ventrolateral prefrontal cortex during cognitive control execution to previously learned high gain relative to low gain cues. Older adolescents also expressed increased coupling between the dorsal striatum and dorsolateral prefrontal cortex for high gain cues, whereas younger adolescents expressed increased coupling between the striatum and ventromedial prefrontal cortex. These findings reveal that learned high value cue-incentive associations enhance cognitive control in late adolescence in parallel with value-selective recruitment of corticostriatal systems.

Contributions of dopamine-related basal ganglia neurophysiology to the developmental effects of incentives on inhibitory control

“…The age-specificity of this effect, evidenced by both a significant age by incentive interaction and the difference in developmental trajectories between incentive conditions ( Fig. 3 B), suggests that adolescents may be particularly sensitive to the enhancing effects of rewards, in line with previous findings in adolescent cognitive control ( Geier et al, 2010 , Insel et al, 2019 , Insel et al, 2017 , Padmanabhan et al, 2011 , Teslovich et al, 2014 , Van Duijvenvoorde et al, 2016 but see Rodman et al, 2021 who showed that rewards similarly modulated effort in adolescents and adults during a physical force task). This increased reward sensitivity may be driven by developmental changes within the Daergic system during adolescence ( Andersen et al, 1997 , Luciana et al, 2012 , Padmanabhan and Luna, 2014 , Wahlstrom et al, 2010 ) and increased reward drive (i.e., increased motivation when rewards are at stake) that facilitates exploration and experience-dependent plasticity ( Luciana et al, 2012 ), required for specialization of frontostriatal circuits and refinements in cognitive processes into adulthood.…”

Friendship stability in adolescence is associated with ventral striatum responses to vicarious rewards

“…We acknowledge the shortcomings of using losing as a baseline condition when examining vicarious reward responses of the ventral striatum, although winning vs. losing is a commonly used contrast to examine reward sensitivity (for self; e.g., refs. 6,12,38,39 ). In an attempt to examine reward sensitivity (i.e., when winning) and punishment sensitivity (i.e., when losing) in isolation, we used the self-reported pleasure ratings to conduct post-hoc analyses.…”

Reward enhances memory via age-varying online and offline neural mechanisms across development

“…Eighty-nine participants ages 8- to 25-years-old (M age = 16.16, SD age = 4.67, 45 female) were included in analyses. A target sample size of n = 90, including 30 children, 30 adolescents, and 30 adults, was determined based on prior work using similar or smaller sample sizes to identify age-related differences in behavior and brain activation (Van Den Bos and Rodriguez, 2015; Insel et al, 2019; Callaghan et al, 2021). Data exclusions consisted of: eight participants with excessive motion (participants without at least one complete encoding, baseline arrows, and post-encoding arrows runs due to exclusions of runs with 15% or more timepoints censored with greater than 0.9 mm relative translational motion), seven participants who elected to not complete or terminate the fMRI scan, and five participants with incomplete datasets as a result of fMRI scanner malfunction.…”