Characterizing decision-making and reward processing in bipolar disorder: A cluster analysis

Abstract

The presence of abnormalities in emotional decision-making and reward processing among bipolar patients (BP) has been well rehearsed. These disturbances are not limited to acute phases and are common even during remission. In recent years, the existence of discrete cognitive profiles in this psychiatric population has been replicated. However, emotional decision making and reward processing domains have barely been studied. Therefore, our aim was to explore the existence of different profiles on the aforementioned cognitive dimensions in BP. The sample consisted of 126 euthymic BP. Main sociodemographic, clinical, functioning, and neurocognitive variables were gathered. A hierarchical-clustering technique was used to identify discrete neurocognitive profiles based on the performance in the Iowa Gambling Task. Afterward, the resulting clusters were compared using ANOVA or Chi-squared Test, as appropriate. Evidence for the existence of three different profiles was provided. Cluster 1 was mainly characterized by poor decision ability. Cluster 2 presented the lowest sensitivity to punishment. Finally, cluster 3 presented the best decision-making ability and the highest levels of punishment sensitivity. Comparison between the three clusters indicated that cluster 2 was the most functionally impaired group. The poorest outcomes in attention, executive function domains, and social cognition were also observed within the same group. In conclusion, similarly to that observed in "cold cognitive" domains, our results suggest the existence of three discrete cognitive profiles concerning emotional decision making and reward processing. Amongst all the indexes explored, low punishment sensitivity emerge as a potential correlate of poorer cognitive and functional outcomes in bipolar disorder.