The Role Of The Ventromedial Prefrontal Cortex In Value-Based Decision-Making

The ventromedial prefrontal cortex (vmPFC) has been shown to correlate with the subjective value for options, across reward type and across hundreds of functional neuroimaging studies. Despite the prominence of its role in preference-based decision-making, its specific contributions to how decisions are made have not yet been well-characterised. Study 1 addresses what the vmPFC signal represents during decision-making. While the vmPFC signal has been shown to correlate highly with subjective value in past studies, this signal is also consistent with mental navigation through a conceptual attribute space using a grid-like code. We found that the mental navigation model lacked support in the evidence, and the subjective value model remains the best explanation for vmPFC signal during decision-making. After having established that the signal in vmPFC reflects subjective value, Study 2 addresses whether subjective value representations remain consistent for non-choice preference tasks, and when this representation comes online during the decision process. This study shows that the value network seen previously for choice tasks also is active during a matching bidding task, and that the vmPFC, interestingly, represents value only at the time of the final choice. Finally, in Study 3, I address the question of how the vmPFC is necessary for subjective value in my third chapter. Transitivity (the idea that if A \u3e B, and B \u3e C, then A \u3e C) is a key property of a value-based system. Individuals with ventromedial frontal lobe damage have been found to make more transitivity errors in the past, but it is not known whether vmPFC damage causes fundamentally intransitive choices (implying abolishment of value), or transitive but noisier choices (implying preservation of value but increased instability). We found strong evidence for the second case, demonstrating that vmPFC damage adds instability to valuation but does not abolish it. The evidence I present here is consistent with the theory that vmPFC is involved in a subjective value-based process during decision-making, yet that value is a distributed process over many brain regions where other regions may compensate for the loss of the vmPFC in calculating value