Abnormal responses to monetary outcomes in cortex, but not in the basal ganglia, in schizophrenia

Abstract

Psychosis has been associated with aberrant brain activity concurrent with both the anticipation and integration of monetary outcomes. The extent to which abnormal reward-related neural signals can be observed in chronic, medicated patients with schizophrenia (SZ), however, is not clear. In an fMRI study involving 17 chronic outpatients with SZ and 17 matched controls, we used a monetary incentive delay (MID) task, in which different-colored shapes predicted gains, losses, or neutral outcomes. Subjects needed to respond to a target within a time window in order to receive the indicated gain or avoid the indicated loss. Group differences in blood-oxygen-level-dependent responses to cues and outcomes were assessed through voxel-wise whole-brain analyses and regions-of-interest analyses in the neostriatum and prefrontal cortex (PFC). Significant group by outcome valence interactions were observed in the medial and lateral PFC, lateral temporal cortex, and amygdalae, such that controls, but not patients, showed greater activation for gains, relative to losses. In the striatum, neural activity was modulated by outcome magnitude in both groups. Additionally, we found that ratings of negative symptoms in patients correlated with sensitivity to obtained losses in medial PFC, obtained gains in lateral PFC, and anticipated gains in left ventral striatum. Sensitivity to obtained gains in lateral PFC also correlated with positive symptom scores in patients. Our findings of systematic relationships between clinical symptoms and neural responses to stimuli associated with rewards and punishments offer promise that reward-related neural responses may provide sensitive probes of the effectiveness of treatments for negative symptoms.

Figures

Figure 1

Schematic of a single trial from the experimental task. The subject was first presented with a single valence cue (eg, the ‘win' cue, circled in red) lasting 250 ms. Following an ISI of 0.5–3.5 s, the subject was then presented with a single value cue (eg, the $15 value cue, circled in red; there were three separate values of anticipated gains, three separate values of anticipated losses, and one neutral value) lasting 350 ms. For gain trials, the magnitude cue consisted of scanned images of exact dollar amounts ($2.50–15), whereas for loss trials the magnitude cue consisted of scanned images of exact dollar amounts ($1.50–9) with a large red ‘X' superimposed. Following another ISI of 0.5–3.5 s, a target stimulus (a white cross on a black background) appeared, prompting the participant to respond within the target time window. The target display time and corresponding response window ranged from 250 to 500 ms as a function of the participant's recent response history, and was followed by an ISI of 150–400 ms. Following this ISI, the actual outcome was displayed for 1 s. If a participant responded within the acceptable response window on a gain trial, the total amount of money incremented by the amount in the magnitude cue ($2.50–15.00); if not, the total incremented by $1.25 (half of the smallest gain magnitude cue). If a participant responded within the acceptable response window on a loss trial, the total amount of money decremented by 75 cents (half of the smallest loss magnitude cue); if not, the total decremented by the amount in the magnitude cue ($1.50–9.00).

Figure 2

ROI analyses of reward anticipation responses. (a) ROIs in left (yellow; −8, 12, 0) and right (red; 11, 11, 0) ventral striatum (VS), based on the work of Knutson et al (2001). (b) Scatter plot showing correlation between patients' summed avolition and anhedonia scores from the SANS and activity related to anticipated gains in left VS. Patients with higher ratings of negative symptoms showed more blunted activity related to anticipated gains.

Figure 3

Brain regions showing significant group × outcome valence interactions. In all areas, controls showed more positive responses to gains than to losses. Patients, by contrast, showed either similar activations to large gains and losses, or reduced responses to gains, relative to losses. Color scale shows voxel-wise F-value, with functional images thresholded at F=11.5 (p=0.001). (a) BOLD responses to negative and positive outcomes by patients (red lines) and controls (blue lines) shown in medial PFC/pregenual ACC (region 1, Talairach coordinates: −1, 41, 5). A similar pattern was observed in the left precentral gyrus (2: −50, −3, −44). Anatomical image cut at x=3, y=−7. (b) BOLD responses by patients and controls shown in the right inferior frontal gyrus (3: 50, 43, 4). A similar pattern was observed in the right superior frontal gyrus (4: 29, 43, 38). Anatomical image cut at y=41. (c) BOLD responses by patients and controls shown in the left amygdala (8: −23, −13, −15). Similar patterns were observed in the right amygdala (7) and middle temporal gyrus/BA 21, bilaterally (5 and 6; see Table 2 for coordinates). Anatomical image cut at y=−11, z=−7.

Figure 4

Brain regions showing main effects of outcome valence. In all areas, the entire sample of subjects showed more positive responses to gains than to losses. Color scale shows voxel-wise F-value, with functional images thresholded at F=11.5 (p=0.001). (a) 1: right ventral striatum (Talairach coordinates: 14, 3, −5); 2: right middle frontal gyrus/BA 8 (47, 11, 38); 3: left precentral gyrus, BA 6 (−51, 3, 34). Anatomical image cut at y=3. (b) 4: dorsal ACC (−6, 29, 42). Anatomical image cut at x=−5. (c) 5: right insula (38, −1, 12); 6: right caudate (12, 17, 7). Anatomical image cut at y=−3 (right of x=22) and y=19 (left of x=22). (d) Four regions of PFC. 7: right BA 46 (57, 24, 24); 8: right BA 8 (6, 52, 36); 9: left BA 8 (−13, 56, 43); 10: left BA 46 (−46, 40, 31). Anatomical image cut at y=24 (right of x=25) and y=54 (left of x=25). (e) BOLD responses to negative and positive outcomes by patients (red lines) and controls (blue lines) shown in the right ventral striatum (region 1). (f) BOLD responses by patients and controls shown in the right insula (region 5).

Figure 5

BOLD responses in SZ patients and controls in the striatum, bilaterally, to large losses (hatched red; $9, $6, $1.50), successfully minimized losses (solid red; $0.75), minimum gains (hatched green; $1.25), and successfully maximized gains (solid green; $15, $10, $2.50). Both patients and controls showed differentiation between successful and unsuccessful trials in the striatum, bilaterally. Talairach coordinates in Table 4. Color scale shows voxel-wise F-value, with functional images thresholded at F=11.5 (p=0.001).