Abstract
In human decision-making under risk, loss is typically valued more than the same amount of gain, a behavioral phenomenon known as loss aversion, which suggests that gain and loss are evaluated differently in the brain. Most previous neuroimaging studies focused on the brain regions that show differential responses to losses relative to gains. What is still largely unknown is how the neural processing of gain and loss may unfold in time and drives loss aversion. Here, we designed a gambling task ideal for investigating the temporal course of the valuation process and used magnetoencephalography (MEG) to track human participants' brain activities for valuating gain and loss. Computational modeling of participants' behaviors implies that the gain and loss presented simultaneously can compete for cognitive resources, during which loss signals dominate the valuation process, resulting in loss aversion. Indeed, time-resolved MEG analysis reveals that the evaluation process of loss terminated later for participants with higher loss aversion than those with lower loss aversion, though the gain valuation had similar temporal courses for different participants. These results suggest that the origin of loss aversion may lie in the neural dynamics of loss processing.