Amygdala and nucleus accumbens in responses to receipt and omission of gains in adults and adolescents

Abstract

Adolescents' propensity for risk-taking and reward-seeking behaviors suggests a heightened sensitivity for reward, reflected by greater feedback-related activity changes in reward circuitry (e.g., nucleus accumbens), and/or a lower sensitivity to potential harm reflected by weaker feedback-related activity changes in avoidance circuitry (e.g., amygdala) relative to adults. Responses of nucleus accumbens and amygdala to valenced outcomes (reward receipt and reward omission) were assayed using an event-related functional magnetic resonance imaging procedure paired with a monetary reward task in 14 adults and 16 adolescents. Bilateral amygdala and nucleus accumbens showed significantly greater activation when winning than when failing to win in both groups. Group comparisons revealed stronger activation of left nucleus accumbens by adolescents, and of left amygdala by adults. When examining responses to reward receipts and to reward omissions separately, the most robust group difference was within the amygdala during reward omission. The reduction of the fMRI BOLD signal in the amygdala in response to reward omission was larger for adults than for adolescents. Correlations showed a close link between negative emotion and amygdala decreased BOLD signal in adults, and between positive emotion and nucleus accumbens activation in adolescents. Overall, these findings support the notion that the signal differences between positive and negative outcomes involve the nucleus accumbens more in adolescents than in adults, and the amygdala more in adults than in adolescents. These developmental differences, if replicated, may have important implications for the development of early-onset disorders of emotion and motivation.

Introduction

Adolescence is a critical period characterized by high levels of risk-taking, impulsivity, and novelty-seeking (Arnett, 1999). Adolescents present vulnerability for gambling (Chambers and Potenza, 2003) and addiction (Chambers et al., 2003). Their propensity for risk-taking may relate to increased responses to reward (Laviola et al., 2003), and reduced harm avoidance (Maggs et al., 1995, Wills et al., 1994). Consummatory processing (responses to outcomes) influences the motivation to select specific courses of actions, and subsequently can indirectly affect pattern of decision-making.

A growing body of work in animals is being conducted to better understand the biological basis of reward-related processing. Single-cell recording studies reveal that dopamine neurons in non-human primates are activated at the receipt of unpredicted rewards, emit no signal after a predicted reward, and emit a negative signal (decreased spike production from baseline) when a predicted reward fails to occur (Schultz et al., 1997).

Similarly, neuroimaging studies in human adults (Breiter et al., 2001, Delgado et al., 2000, Knutson et al., 2001, O'Doherty et al., 2001) reveal that reward receipt enhances activity of subcortical structures (i.e., striatum and amygdala), and of cortical regions (i.e., mesial prefrontal cortex), whereas reward omission, or exposure to losses, tends to reduce the fMRI BOLD signal in these same structures.

At present, two neuroimaging studies have examined reward circuitry in adolescents (Bjork et al., 2004, May et al., 2004), only one with an adult comparison group (Bjork et al., 2004). Bjork et al. (2004) found less activation of the right nucleus accumbens (NAcc) during anticipation of gains in adolescents than in adults. They interpreted this finding as indicating a selectively weaker recruitment of motivational, but not consummatory, components of reward-directed behavior in adolescents. Here, we propose that age-related differences in consummatory behavior and underlying neural substrates may contribute to risk-taking in adolescence.

Whereas increased risk-taking in adolescents can relate to greater reward responses, through enhanced appetitive value of positive outcomes (Arnett, 1992, Gardner and Herman, 1991, Laviola et al., 2003), it can also indicate decreased response to danger or threat, through a decreased aversiveness of negative outcomes, leading to reduced harm avoidance behavior (Maggs et al., 1995, Wills et al., 1994, Wilson and Daly, 1985). The amygdala circuitry is critically involved in anticipation of negative outcomes but also in the response to negative stimuli (LeDoux, 2000). Although participating in the emotional coding of both appetitive and aversive stimuli (Baxter and Murray, 2002, Zald, 2003), the amygdala can function as a behavioral “brake” (Zald, 2003). Its functional characteristics during adolescence are virtually unknown (Killgore et al., 2001, Yurgelun-Todd et al., 2003), although neuroimaging studies suggest distinct patterns of amygdala modulation in affective paradigms during adolescence and adulthood (Monk et al., 2003). These studies have focused on various aspects of the amygdala, including its modulation by attention during processing of fearful facial expression (Monk et al., 2003, Thomas et al., 2001), and its response to passive viewing of fearful and neutral faces (Thomas et al., 2001). Adolescents seemed to be more sensitive to the emotional content than to the attentional demands of the task in Monk et al.'s (2003) study, and to be more responsive to neutral vs. fearful expressions than adults in Thomas et al.'s (2001) study. None of these studies have examined reward-related processes per se.

Here, we test whether adults and adolescents differ in the NAcc's and amygdala's involvement in reward-related processes, specifically in response to reinforcers. Converging data emerge from three different perspectives: the approach/avoidance model of adolescent behavior (Chambers et al., 2003, Spear, 2000), the available neuroimaging findings in adults (Breiter et al., 2001, Delgado et al., 2000, Knutson et al., 2001, O'Doherty et al., 2001), and the electrophysiological work in primates (Schultz, 2002, Schultz et al., 1997). Taken together, this work generates the hypotheses that (1) responses to outcome will affect the NAcc more in adolescents than in adults, and (2) the amygdala more in adults than in adolescents, and (3) that self-ratings of affective responses will correlate with regional activation.