Developmental effects of decision-making on sensitivity to reward: an fMRI study

Abstract

Studies comparing neural correlates of reward processing across development yield inconsistent findings. This challenges theories characterizing adolescents as globally hypo- or hypersensitive to rewards. Developmental differences in reward sensitivity may fluctuate based on reward magnitude, and on whether rewards require decision-making. We examined whether these factors modulate developmental differences in neural response during reward anticipation and/or receipt in 26 adolescents (14.05±2.37 yrs) and 26 adults (31.25±8.23 yrs). Brain activity was assessed with fMRI during reward anticipation, when subjects made responses with-vs.-without decision-making, to obtain large-vs.-small rewards, and during reward receipt. When reward-receipt required decision-making, neural activity did not differ by age. However, when reward receipt did not require decision-making, neural activity varied by development, reward magnitude, and stage of the reward task. During anticipation, adolescents, but not adults, exhibited greater activity in the insula, extending into putamen, and cingulate gyrus for large-vs.-small incentives. During feedback, adults, but not adolescents, exhibited greater activity in the precuneus for large-vs.-small incentives. These data indicate that age-related differences in reward sensitivity cannot be characterized by global hypo- or hyper-responsivity. Instead, neural responding in striatum, prefrontal cortex and precuneus is influenced by both situational demands and developmental factors. This suggests nuanced maturational effects in adolescent reward sensitivity.

Fig. 1

Reward task design; results of self-report and behavioral data. (A) Each trial was comprised of reward anticipation (1500 ms) and receipt (1000 ms) stages. During reward anticipation, colored circle cues prompted subjects to respond. Cue color indicated the type of trial (no-choice: orange; choice: blue; motor: yellow). Size of no-choice and choice trials cues reflected incentive size (small: 3-points; large: 6-points). For no-choice trials, a “1” or “2” within the cue directed subjects to press the corresponding button to gain points. For choice trials, a “?” within the cue prompted subjects to guess the correct response to gain points. For motor trials, a small blank cue prompted subjects to press any button, with no opportunity to gain points. During reward receipt, subjects learned if their response was rewarded and their cumulative score. Incidental stimuli (1000 ms) were purple rectangles that introduced variable-duration jitter into the design, and required no response. Cue color was counterbalanced across subjects. Graphs depict average (±SEM) for (B) post-scan affective ratings, and (C) response times, for no-choice (orange) and choice (blue) trials.

Fig. 2

Reward anticipation. (A) 3-way (age group x decision-making x incentive size) interaction found in insula extending to putamen, and dACC (MNI X-plane = 7; Z-plane = −4). Graphs depict this interaction. Bars represent mean (±SEM) percent signal change for large versus small incentives, during no-choice (orange) and choice (blue) reward anticipation in (B) right insula extending to striatum, and (C) right dACC.

Fig. 3

Reward receipt. (A) 3-way (Age X Decision-making X Incentive size) interaction found in precuneus (MNI X-plane = −37). Graph depict this interaction. Bars represent mean (±SEM) percent signal change for large versus small incentives, during no-choice (orange) and choice (blue) reward receipt in (B) left precuneus.