Increased Ventromedial Prefrontal Cortex Activity in Adolescence Benefits Prosocial Reinforcement Learning

Abstract

Learning which of our behaviors benefit others contributes to forming social relationships. An important period for the development of (pro)social behavior is adolescence, which is characterized by transitions in social connections. It is, however, unknown how learning to benefit others develops across adolescence and what the underlying cognitive and neural mechanisms are. In this functional neuroimaging study, we assessed learning for self and others (i.e., prosocial learning) and the concurring neural tracking of prediction errors across adolescence (ages 9-21, N = 74). Participants performed a two-choice probabilistic reinforcement learning task in which outcomes resulted in monetary consequences for themselves, an unknown other, or no one. Participants from all ages were able to learn for themselves and others, but learning for others showed a more protracted developmental trajectory. Prediction errors for self were observed in the ventral striatum and showed no age-related differences. However, prediction error coding for others showed an age-related increase in the ventromedial prefrontal cortex. These results reveal insights into the computational mechanisms of learning for others across adolescence, and highlight that learning for self and others show different age-related patterns.

Keywords: Adolescence; Cognitive empathy; Prediction error; Prosocial; Reinforcement learning; Ventromedial prefrontal cortex.

Fig. 1

Prosocial learning task and behavioral data. (A) Participants played a two-choice probabilistic reinforcement learning task in which outcomes resulted in monetary consequences for themselves (Self condition), for an unknown other participant in the experiment who could not reciprocate (Other condition), or for No One. (B) Group-level performance across trials (learning curves) per condition, averaged across blocks. Performance represents the fraction of selecting the stimulus with a high reward contingency. The dashed line indicates performance at chance level (0.5). (C) Performance per condition per age cohort, averaged across the entire task. In all conditions, performance improved across trials, but an age-related increase was only observed when learning for others. Note that age is used as a continuous variable in all analyses but is visualized as age cohorts for illustrative purposes. The age-related increase was greater for the Other than for the Self and No One condition. (D) Learning rates per condition per age cohort. Age-related decreases in learning rates are only observed in the Self and Other condition. The age-related decrease in learning rate was greater in the Other compared to the Self and No One condition. Asterisks indicate significant effects. Error bars represent standard error of the mean (s.e.m.).

Fig. 2

Common prediction error (PE) coding in three regions of interest. Shown are the responses to prediction errors for Self, Other, and No One in (A) the vmPFC, (B) left sgACC, and (C) ventral striatum. (D) Significant clusters of activation in the vmPFC (blue), sgACC (cyan), and ventral striatum (yellow). All images displayed at p < .05 FWE-SVC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Ventral striatum response to prediction errors for Self versus Other. (A) Left ventral striatum [x = 12, y = 11, z = − 11] response for Self PE and Other PE. (B) Overlay of the response for Self PE > Other PE in the left ventral striatum. All images displayed at p < .05 FWE-SVC.

Fig. 4

Linear age effects in responses to Other PE > Self PE in the vmPFC. (A) scatterplot showing the relation between age and activation in the vmPFC for Other PE > Self PE. Scatterplot is only presented for visualization. (B) Overlay of the response for Other PE > Self PE in the vmPFC [− 15, 41, − 11]. All images displayed at p < .05 FWE-SVC.

Fig. 5

Relation of cognitive empathy with performance for Others and learning rate for Others. (A) Partial correlation plot showing that individuals with more cognitive empathy perform better for Others (controlled for performance for Self). (B) Partial correlation plot showing that individuals with more cognitive empathy have lower learning rates when learning for Others (controlled for learning rate for Self).