Cognitive flexibility in adolescence: neural and behavioral mechanisms of reward prediction error processing in adaptive decision making during development

Abstract

Adolescence is associated with quickly changing environmental demands which require excellent adaptive skills and high cognitive flexibility. Feedback-guided adaptive learning and cognitive flexibility are driven by reward prediction error (RPE) signals, which indicate the accuracy of expectations and can be estimated using computational models. Despite the importance of cognitive flexibility during adolescence, only little is known about how RPE processing in cognitive flexibility deviates between adolescence and adulthood. In this study, we investigated the developmental aspects of cognitive flexibility by means of computational models and functional magnetic resonance imaging (fMRI). We compared the neural and behavioral correlates of cognitive flexibility in healthy adolescents (12-16years) to adults performing a probabilistic reversal learning task. Using a modified risk-sensitive reinforcement learning model, we found that adolescents learned faster from negative RPEs than adults. The fMRI analysis revealed that within the RPE network, the adolescents had a significantly altered RPE-response in the anterior insula. This effect seemed to be mainly driven by increased responses to negative prediction errors. In summary, our findings indicate that decision making in adolescence goes beyond merely increased reward-seeking behavior and provides a developmental perspective to the behavioral and neural mechanisms underlying cognitive flexibility in the context of reinforcement learning.

Keywords: Adolescence; Cognitive flexibility; Development; Functional magnetic resonance imaging (fMRI); Reward prediction errors.

Fig. 1

Probabilistic reversal learning task. On each trial (average duration: 9000 ms), two stimuli were simultaneously presented. The participant had to select one of the stimuli within 1500 ms. The selected stimulus was highlighted until the end of the stimulus presentation (2500 ms). After a jittered interstimulus interval (2000–4000 ms), the outcome was displayed for 1000 ms. Rewards were indicated by a framed coin whereas punishments were depicted by a crossed coin. Between trials, a jittered fixation cross was shown (2000–4000 ms).

Fig. 2

Learning rate differences between adolescents and adults. The parameters from the RSAV model show an increased learning rate for negative RPEs in chosen stimuli (α_c⁻). The other learning rates did not significantly differ. *: p < .05, multiple comparison corrected.

Fig. 3

Differences between adolescents and adults in the RPE network. (A) A network containing the dmPFC (upper panel) and the aIns (lower panel) shows increased activation for decreasing RPEs among all subjects. (B) A group comparison between the adolescents and adults reveals a significant activation difference in the right aIns. (C) Subsequent exploratory analysis revealed that this group difference was mainly driven by an increased activation for negative RPEs in adolescents. ***: p < .001.