Previously reward-associated stimuli persistently capture attention. We attempted to extinguish this attentional bias through a reversal learning procedure where the high-value color changed unexpectedly. Attentional priority shifted during training in favor of the currently high-value color, although a residual bias toward the original high-value…