Brain mechanism underlying Iowa gambling task : an fMRI study

Abstract

This study employed event-related functional Magnetic Resonance Imaging (fMRI) to measure the brain activation while performing the Iowa Gambling Task (IGT). The purpose was to explore the brain mechanism underlying decision-making process in ambiguous conditions, and has implications for impulsive behavior assessment and intervention. The IGT simulates decision-making process under uncertain conditions. The neural process underlying the IGT was explained by the Somatic Marker Hypothesis (SMH), which described how the somatic signals generated from the body were represented and regulated in the brain, and how they then influenced the decision-making process. According to the hypothesis, a neural circuitry involving emotion, memory and behavioral system was engaged. This neural circuitry was generally supported by lesion studies and neuroimaging studies. With event-related fMRI, it is possible to indentify distinct underlying neural substrates of the temporally separated task components involved in the IGT task. However, existing fMRI studies employing the IGT task are still relatively scarce and provide inconsistency results, partly due to the complexity of the task and the different control tasks employed by the studies. Furthermore, most fMRI studies only concerned the main contrast between risky and safe decks of cards but ignored the effect of another factor in the task design, the punishment frequency associated with the decks of cards, which may influenced subjects' preference on decks of cards. This study explored the brain activation using fMRI during the performance of the original design of the IGT with 2 by 2 factorial design, and aimed to uncover: 1) the brain activation related to the long-term outcome effect, i.e. the risky versus safe contrast; 2) whether this contrast would be influenced by the factor of punishment frequency; and 3) the brain activation related to the learning process in IGT. The behavioral results showed that the subjects could gradually learn to choose more from the safe decks of cards and avoid the risky decks of cards during the task, and that they were more sensitive to the risky versus safe contrast under the condition with a higher punishment frequency. The neuroimaging results demonstrated that choices from the risky decks produced higher activation in anterior cingulate cortex (ACC), and this contrast was dominated by the one in decks of cards with a higher punishment frequency. The learning process was assessed by comparing brain activation during choices in the first half trials with those in the second half trials, which detected higher activation in insula, amygdala and hippocampus regions during the earlier phase of the task. The results generally reflected the SMH framework: behavioral performance was related to the cortical region for the somatic states to influence the selection behaviors; and the time effect was reflected in the brain regions crucial to the generation and the cortical representation of the somatic state.