Alcohol Dependence and Altered Engagement of Brain Networks in Risky Decisions

Alcohol dependence is associated with heightened risk tolerance and altered decision-making. This raises the question as to whether alcohol dependent patients (ADP) are incapable of proper risk assessment. We investigated how healthy controls (HC) and ADP engage neural networks to cope with the increased cognitive demands of risky decisions. We collected fMRI data while 34 HC and 16 ADP played a game that included “safe” and “risky” trials. In safe trials, participants accrued money at no risk of a penalty. In risky trials, reward and risk simultaneously increased as participants were instructed to decide when to stop a reward accrual period. If the participant failed to stop before an undisclosed time, the trial would “bust” and participants would not earn the money from that trial. Independent Component Analysis was used to identify networks engaged during the anticipation and the decision execution of risky compared with safe trials. Like HC, ADP demonstrated distinct network engagement for safe and risky trials at anticipation. However, at decision execution, ADP exhibited severely reduced discrimination in network engagement between safe and risky trials. Although ADP behaviorally responded to risk they failed to appropriately modify network engagement as the decision continued, leading ADP to assume similar network engagement regardless of risk prospects. This may reflect disorganized network switching and a facile response strategy uniformly adopted by ADP across risk conditions. We propose that aberrant salience network (SN) engagement in ADP might contribute to ineffective network switching and that the role of the SN in risky decisions warrants further investigation.