Decision making and Learning
The power of belief and how it affects decision-making
Human decision-making has been shown to involve "trial-and-error" processes (aka reinforcement learning). These RL signals are largely dependent on dopaminergic brain regions such as the striatum. Many neuroactive drugs such as nicotine interfere with these DA signals. While many studies have suggested that DA-dependent RL processes are "hard-wired", recent findings from our lab suggest that belief could have a powerful effect on these neurophysiological signals. In a group of chronic smoker, we found that the instructed belief of “no nicotine in cigarette” (compared with “nicotine in cigarette”) strongly reduced neural responses in the striatum to RL signals (value, prediction error) and reduced the impact of both on smokers’ choices. Thus, belief is a powerful mechanism through which drug-driven physiological signals can be overwritten.
Gu X, Lohrenz T, Salas R, Baldwin PR, Soltani A, Kirk U, Cinciripini PM, Montague PR (2015) Belief about nicotine selectively modulates value and reward prediction error signals in smokers. Proceedings of the National Academy of Sciences.
Volkow ND, Baler R (2015) Beliefs modulate the effects of drugs on the human brain. Proc Natl Acad Sci U S A 112:2301-2302.
Emotion regulation of "fictive" decision signals
Computational models of reward processing suggest that foregone or fictive outcomes (“what might have happened” or "regret") serve as important information sources for learning. An outstanding question is how these learning signals interact with top-down cognitive influences, such as cognitive reappraisal strategies. Using a sequential investment task and fMRI, we show that reappraisal, a well-established emotion regulation strategy, significantly attenuates the influence of fictive, but not experiential learning signals on investment behavior in healthy adults. On the neural level, fictive learning signals are modulated by reappraisal resulting in changes in neural activity and connectivity in the AIC. In summary, these results contribute novel evidence by demonstrating that individual differences are targeted by differential learning signals highlighting the importance of adapting cognitive reappraisal strategies based on individual susceptibility in treatment settings in cases such as addiction.
Gu, X., Kirk, U., Lohrenz, T.M. & Montague, P.R. Cognitive strategies regulate fictive, but not reward prediction error signals in a sequential investment task. Hum Brain Mapp (2013). [link]
Differential neural systems supporting information updating and valuation during decision-making
Previous neuroimaging studies have highlighted the involvement of both the insular and ventromedial prefrontal (vmPFC) cortices in decision-making. However, the necessity and dissociability of their involvement remain unclear. Using model-based computational modelling and neuropsychological lesion approaches, we found that patients with insula lesions displayed abnormally low learning rate, while patients with vmPFC lesions displayed reduced sensitivity to fairness. These findings provide compelling computational and lesion evidence supporting the necessary, yet dissociable roles of the insula and vmPFC in decision-making in humans.
Gu X, Wang X, Hula A, Wang S, Xu S, Lohrenz TM, Knight RT, Gao Z, Dayan P, Montague PR. Necessary, Yet Dissociable Contributions of the Insular and Ventromedial Prefrontal Cortices to Norm Adaptation: Computational and Lesion Evidence in Humans. J Neurosci 35:467-473.