Neural function underlying reward expectancy and attainment in adolescents with diverse psychiatric symptoms

Abstract

Reward dysfunction has been hypothesized to play a key role in the development of psychiatric conditions during adolescence. To help capture the complexity of reward function in youth, we used the Reward Flanker fMRI Task, which enabled us to examine neural activity during expectancy and attainment of both certain and uncertain rewards. Participants were 84 psychotropic-medication-free adolescents, including 67 with diverse psychiatric conditions and 17 healthy controls. Functional MRI used high-resolution acquisition and high-fidelity processing techniques modeled after the Human Connectome Project. Analyses examined neural activation during reward expectancy and attainment, and their associations with clinical measures of depression, anxiety, and anhedonia severity, with results controlled for family-wise errors using non-parametric permutation tests. As anticipated, reward expectancy activated regions within the fronto-striatal reward network, thalamus, occipital lobe, superior parietal lobule, temporoparietal junction, and cerebellum. Unexpectedly, however, reward attainment was marked by widespread deactivation in many of these same regions, which we further explored using cosine similarity analysis. Across all subjects, striatum and thalamus activation during reward expectancy negatively correlated with anxiety severity, while activation in numerous cortical and subcortical regions during reward attainment positively correlated with both anxiety and depression severity. These findings highlight the complexity and dynamic nature of neural reward processing in youth.

Keywords: Adolescence; Anxiety; Depression; Reward attainment; Reward expectancy; fMRI.

Fig. 1

Reward Flanker Task (RFT) includes: 1) Cues indicating high (50¢), low (10¢), no (0¢), or uncertain (?) reward value; 2) Flanker stimulus; 3) Calibrated window to respond to flanker stimulus.;4) Reward or non-reward feedback.

Fig. 2

Neural activation maps (p_TFCE-FWE < 0.05) during a) reward expectancy (reward vs. non-reward cues) and b) reward attainment (reward vs non-reward outcomes after both certain and uncertain cues). The background overlay is sulcus depth averaged from HCP 1200 subjects dataset. Black lines on the surface represent HCP MMP Atlas (Glasser et al., 2016).

Fig. 3

Neural activation during reward attainment was positively associated with the severity of a) depression (CDRS-R) and b) anxiety (MASC) symptoms. c) Neural activation during reward expectancy was negatively associated with anxiety severity. Maps display Pearson’s r values thresholded at p_TFCE-FWE < 0.05. The background overlay is mean sulcus depth from the HCP 1200 subject dataset. Black lines on the surface represent HCP MMP Atlas.