Neural correlates of attachment in adolescents with trauma: a preliminary study on frustrative non-reward

Abstract

Despite the proposed early life origins of attachment style and its implications for risk for psychopathology, little is known about its neurodevelopmental course. Adolescence represents a key transition period when neural substrates of emotion regulation and reward undergo dramatic maturational shifts. Thus, maladaptive coping strategies associated with insecure attachment styles may have an exaggerated effect during adolescence. The current study, therefore, examined the neural correlates of insecure attachment in a diverse sample of adolescents using a frustrative non-reward task (i.e. repeatedly being denied an expected reward). Although there were no significant interactions in the whole-brain activation averaged over the course of the task, the use of complementary analytic approaches (connectivity, change in activation over the course of the task) revealed widespread alterations associated with avoidant attachment during the immediate reaction to, and ensuing recovery from, being denied a reward. Most strikingly, increased avoidant attachment, adjusting for anxious attachment, predicted functional connectivity and change in activity over time in amygdala-prefrontal and frontostriatal networks to reward blocked vs received trials. These patterns were in the opposite direction compared to those exhibited by adolescents lower in avoidant attachment. The findings suggest that negative emotional experiences, such as receiving frustrating feedback, may be uniquely aversive internal experiences for avoidantly attached adolescents and provide preliminary evidence that early coping strategies may persist into adolescence in the form of altered emotion- and reward-related neural patterns.

Keywords: adolescence; attachment; brain; emotion regulation; frustrative non-reward.

Fig. 1.

fMRI frustrative non-reward task. In this novel task, a modified version of an adolescent-friendly monetary incentive delay task (Dougherty et al., 2018), participants first saw a cue (piñata dropping for 2000 ms) alerting them that a target (piñata) with a possible reward was about to drop, followed by a variable duration (2500–5500 ms) delay period when participants waited for the target to drop (anticipation period). When the target was presented (initially at 500 ms but adjusted in real-time based on performance), participants were instructed the ‘hit’ the piñata by pressing a button on an MRI response box. After a delay, participants received feedback on whether they received a reward (piñata breaking and stars falling into the basket) or were blocked from a reward (piñata swings away, no stars in basket) (immediate reaction period). Forty percent of hits were followed by positive feedback (reward received), but for 60% of hits, feedback was rigged such that participants were denied a reward (reward blocked). All misses were followed by no reward. After a jittered intertrial interval (2500–5500 ms), the subsequent anticipation period was considered the ensuing recovery period.

Fig. 2.

(A) The avoidant attachment × condition interaction predicts changes in right amygdala connectivity during the immediate reaction period (A.1.1) and the ensuing recovery period (A.1.2). The avoidant attachment × condition interaction predicts changes in right ventral striatum connectivity during the immediate reaction period (A.2.1), and avoidant attachment (main effect) negatively predicts changes in right ventral striatum connectivity during the ensuing recovery period (A.2.2). Avoidant attachment is plotted at 1 s.d. below the mean (‘low’) and 1 s.d. above the mean (‘high’). (B) Avoidant attachment (main effect) predicts changes in brain activation during both the immediate reaction period (B.1) and the ensuing recovery period (B.2). The ‘low avoidant attachment’ line depicts the changes in average brain activation over time for participants with avoidant attachment scores that were 1 s.d., or greater, below the mean; the ‘high avoidant attachment’ line depicts the changes in average brain activation over time for participants with avoidant attachment scores that were 1 s.d., or greater, above the mean. Brain images (threshold set at whole-brain corrected false probability rate of P < 0.05; axial view for all images) depict significant clusters. Only one cluster is plotted as an example when a contrast contains multiple regions with similar patterns.