Tasting rewards. Effects of orosensory sweet signals on human error processing

Human research has shown interactions between rewards and cognitive control. In animal models of affective neuroscience, reward administration typically involves administering orosensory sugar signals (OSS) during caloric-deprived states. We adopted this procedure to investigate neurophysiological mechanisms of reward-cognitive control interactions in humans. We predicted that OSS would affect neurophysiological and behavioral indices of error processing oppositely, depending on the relative weight of the OSS-induced 'wanting' and 'liking' components of reward. We, therefore, conducted a double-blind, non-nutritive sweetener-controlled study with a within-subject design. Fasted (16 hr) participants (N = 61) performed a modified Flanker task to assess neurophysiological (error-related negativity [N_e/ERN]) and behavioral (post-error adaptations) measures of error processing. Non-contingent to task performance, we repeatedly administered either a sugar (glucose) or non-nutritive sweetener (aspartame) solution, which had to be expulsed after short oral stimulation to prevent post-oral effects. Consistent with our hypothesis on how 'liking' would affect N_e/ERN amplitude, we found the latter to be decreased for sugar compared to aspartame. Unexpectedly, we found post-error accuracy, instead of post-error slowing, to be reduced by sugar relative to aspartame. Our findings suggest that OSS may interact with error processing through the 'liking' component of rewards. Adopting our reward-induction procedure (i.e. administering OSS in a state of high reward sensitivity [i.e. fasting], non-contingent to task performance) might help future research investigating the neural underpinnings of reward-cognitive control interactions in humans.