Symptom dimensions are associated with reward processing in unmedicated persons at risk for psychosis

Abstract

There is growing evidence that reward processing is disturbed in schizophrenia. However, it is uncertain whether this dysfunction predates or is secondary to the onset of psychosis. Studying 21 unmedicated persons at risk for psychosis plus 24 healthy controls (HCs) we used a incentive delay paradigm with monetary rewards during functional magnetic resonance imaging. During processing of reward information, at-risk individuals performed similarly well to controls and recruited the same brain areas. However, while anticipating rewards, the high-risk sample exhibited additional activation in the posterior cingulate cortex, and the medio- and superior frontal gyrus, whereas no significant group differences were found after rewards were administered. Importantly, symptom dimensions were differentially associated with anticipation and outcome of the reward. Positive symptoms were correlated with the anticipation signal in the ventral striatum (VS) and the right anterior insula (rAI). Negative symptoms were inversely linked to outcome-related signal within the VS, and depressive symptoms to outcome-related signal within the medial orbitofrontal cortex (mOFC). Our findings provide evidence for a reward-associated dysregulation that can be compensated by recruitment of additional prefrontal areas. We propose that stronger activations within VS and rAI when anticipating a reward reflect abnormal processing of potential future rewards. Moreover, according to the aberrant salience theory of psychosis, this may predispose a person to positive symptoms. Additionally, we report evidence that negative and depressive symptoms are differentially associated with the receipt of a reward, which might demonstrate a broader vulnerability to motivational and affective symptoms in persons at-risk for psychosis.

Keywords: anterior insula; at-risk mental state; dopamine; functional magnetic resonance imaging (fMRI); psychosis; reward; salience processing; ventral striatum.

Figures

Figure 1

Monetary incentive delay task: Example trial and cues representing possible reward outcomes. Participants first saw a cue stipulating with an unpredictable probability the amount of money (4 CHF or 0 CHF) they could win, if they reacted correctly within 750 ms during the ensuing discrimination task, which involved pressing either a left or right button depending upon the direction of a triangle after an anticipation period (variable delay: 2500–3500 ms, mean of 3000 ms). Immediately after target presentation, subjects were informed about the amount of money they had won during this trial and their cumulative total win so far (feedback) for a total of 1500 ms (Abler et al., 2005). The jittered inter-trial interval (ITI) was between 1000 and 8000 ms with a mean of 4000 ms. Trial types were randomly ordered.

Figure 2

Whole-brain group comparison of the contrast reward anticipation vs. no reward anticipation. Subjects at risk for psychosis showed significantly stronger hemodynamic response compared to healthy controls in the posterior cingulate cortex, superior frontal gyrus, and bilateral medio frontal gyrus (corresponding t-values are represented in orange/yellow).

Figure 3

Associations between regions of interest (ROI) and severity of positive symptoms during anticipation of reward. ROIs (depicted in cyan) are overlaid on within-group t-maps for subjects at risk for psychosis (A) for the contrast reward anticipation vs. no reward anticipation (shown in orange, both at a voxel-wise threshold p < 0.001, with an extent of 20 voxels). ROI-based analysis revealed significant association between contrast estimates of the left and right ventral striatum (VS) (A1, A2) and right anterior insula (rAI) (A3) with positive symptom scores (ρ > 0.52, p < 0.015).

Figure 4

Associations among regions of interest (ROIs), clinical symptoms, and reaction time (RT) to cues with possible reward during outcome. ROIs (depicted in cyan) are overlaid on the within-group t-map for subjects at risk for psychosis (A) for the contrast receipt of reward vs. omission of reward (shown in orange, both at a voxel wise threshold of p < 0.001, with an extent of 20 voxels). ROI-based analysis revealed a negative association between contrast estimates within the medio orbitofrontal cortex and severity of depressive symptoms (B1), and the left VS and severity of negative symptoms (B2) (ρ > −0.44, p < 0.045). (C) Signal in the left VS revealed a significant inverse association with RT in healthy controls (blue; ρ = −0.42, p = 0.04) but not for subjects at risk for psychosis (red; ρ = −0.18, p = 0.43).