Our main goal in this study was to compare ventral striatal activity during reward processing in euthymic bipolar I and bipolar II patients and healthy comparison subjects. While all participants exhibited ventral striatal activity during reward anticipation, bipolar II patients showed significantly greater activity than either bipolar I patients or healthy comparison subjects. During positive outcome, however, bipolar I patients showed significantly greater ventral striatal activity than bipolar II patients, but not healthy comparison subjects, although unlike reward anticipation, this difference became nonsignificant after controlling for potential effects of antipsychotic medication. Also supporting our hypotheses, the magnitude of ventral striatal activity during reward anticipation was positively correlated with the fun-seeking scale across all participants. Furthermore, bipolar II patients had enlarged left putamen volume relative to bipolar I patients, and a positive correlation was observed between left putamen volume and left ventral striatal activity during reward anticipation across the whole sample.
The importance of our findings is twofold. First, we demonstrate that bipolar II patients display elevated ventral striatal activity to reward anticipation not only relative to healthy comparison subjects but also relative to bipolar I patients. This difference between bipolar groups was not explained by differences in the severity of residual mood symptoms or antipsychotic medication load. Second, the pattern of ventral striatal activity observed across all participants during reward anticipation was positively associated with the fun/thrill-seeking component of the behavioral activation system (33
). This latter finding concurs with previous research linking ventral striatal activity with the behavioral activation system (34
) and with the dimensional approach to the study of psychiatric disorders advocated by the Research Domain Criteria (17
), which focuses on dimensions of behavior and related neural circuitry function that cut across psychiatric diagnostic categories. Our finding that ventral striatal activity in response to positive outcome did not correlate with any of the behavioral activation system scales is in agreement with Gray’s model (35
), in which this neurobiological system engages in preparation to obtain, but not to receive, reward.
While the two bipolar patient groups and healthy comparison subjects did not differ on self-reported measures of reward sensitivity, bipolar II patients exhibited significantly greater ventral striatal activity during reward anticipation than in other groups. These findings suggest either greater sensitivity of reward neural circuitry activity than self-report measures in differentiating bipolar I and II patients or greater ventral striatal activity for a given level of self-reported fun and thrill seeking during reward anticipation in bipolar II patients than other groups. Either way, elevated ventral striatal activity during reward anticipation may be a potential neural mechanism for greater reward-seeking behavior and mood dysregulation in bipolar II patients. To better understand inconsistencies between findings for self-reported reward sensitivity and ventral striatal activity, future research should include other behavioral measures of reward sensitivity that may be more sensitive to group differences.
Our findings of significantly greater left putamen volume in bipolar II relative to bipolar I patients as well as a positive correlation between left putamen volume and left ventral striatal activity during reward anticipation suggest a structural neural basis for the between-group differences in ventral striatal activity. This finding will require further confirmation given the decrease in statistical significance after controlling for antipsychotic medication load. However, to our knowledge, these results represent the first demonstration of a potential structure-function relationship in reward circuitry in bipolar disorder.
Interestingly, bipolar I patients displayed significantly greater ventral striatal activity than bipolar II patients during positive outcome, suggesting that the former may have been more sensitive to reward receipt than the latter, although this difference did not remain after controlling for the potential effects of antipsychotics. Furthermore, the difference in ventral striatal activity during positive outcome between bipolar I patients and healthy comparison subjects did not meet the cluster-wise significance threshold. Together with the above findings, these may suggest different neural mechanisms for vulnerability to hypomanic or manic episodes in individuals with bipolar I and II disorder—that is, heightened sensitivity to reward anticipation in bipolar II disorder but to reward receipt in bipolar I disorder. Further research would help to clarify this point.
We observed some discrepancies between our results and a previous report (16
) showing significantly greater ventral striatal activity during reward anticipation in euthymic bipolar I patients relative to healthy comparison subjects. Unlike the previous study, participants in our study were not led to believe that their remuneration was dependent on task performance. Instead, we asked participants to perform at their best and told them that performance would be compared across participants. Earlier studies have reported that social competition activates the ventral striatum (36
). Our findings may suggest that the addition of this social reward component may be less effective at activating the ventral striatum than monetary reward during reward anticipation in bipolar I patients.
Exploratory whole-brain analyses revealed left frontotemporal and insula cortical activity during reward anticipation. These neural regions are associated with reward processing, in particular the ventrolateral prefrontal cortex (12
). Between-group comparisons paralleled the pattern of differences in ventral striatal activity: bipolar II patients exhibited greater activity than other groups. While these patterns of between-group differences further support greater sensitivity to reward anticipation in bipolar II than in bipolar I patients or healthy comparison subjects, these results were less robust than the ventral striatal findings. It would therefore be useful to replicate these findings and further clarify to what extent the difference between bipolar patient groups may be related to antipsychotic medication.
Some limitations are noted. The majority of patients with bipolar disorder were receiving psychotropic medications. The potentially confounding effects of psychotropic medication in bipolar neuroimaging research have been discussed previously (37
). Nevertheless, we were able to replicate our findings regarding the differences in ventral striatal activity after controlling for antipsychotic load (in chlorpromazine equivalents) and after excluding patients who were taking this medication class. Our findings in the ventral striatum for reward anticipation appeared to be robust enough to survive these additional analyses, but further replication is warranted, especially because the duration of medication treatment was not recorded during this study. No significant group differences in self-reported measures of reward sensitivity were observed, and our task does not produce a sensitive behavioral output to detect between-group differences in reward sensitivity. Future studies should include other behavioral self-reported measures of reward sensitivity.
In summary, we report significantly elevated ventral striatal activity in bipolar II patients compared with bipolar I patients and healthy comparison subjects, specifically during reward anticipation, and a significant positive correlation between ventral striatal activity to reward anticipation and fun/thrill seeking. These findings suggest that abnormally elevated ventral striatal activity during reward anticipation may be a biomarker of bipolar II disorder, as previously shown for bipolar I disorder, that could ultimately be used to help discriminate bipolar from unipolar depressed patients. Moreover, bipolar I and bipolar II patients may be distinguished by functional differences in ventral striatal activity during reward anticipation and outcome, which in turn may suggest different neural mechanisms predisposing to mood dysregulation across bipolar I and II disorders. Our findings are among the first to identify patterns of abnormal neural activity in bipolar II disorder that may be potential targets for therapeutic interventions for this disabling psychiatric disorder.