How reward value and probability drive human brain function and behavior

Abstract

Humans (as well as animals) have an inherent tendency to seek out rewards and to avoid punishments. This tendency is crucially important in daily life. The brain selects those behavioral actions that have high probabilities of leading to highly valuable outcomes. Therefore, the brain should compute the value as well as the probability of future rewards in order to make optimal decisions. One of the main questions of this thesis was: how does the human brain assess the value and probability of potential future rewards? Other questions that I addressed in this thesis are: How can the subjective value of monetary rewards be assessed? Are there individual differences in reward sensitivity as reflected in learning style? And as reflected in spontaneous fluctuations in brain activity? In Chapter 2 the startle reflex was used as a readout measure of motivational state during the anticipation, obtainment and omission of rewards with varying magnitudes in order to compare the value of these rewards. However, no evidence was found for systematic increases in startle modulation with increased reward magnitude at stake. The aim of Chapter 3 and 4 was to investigate whether there are individual differences in reward sensitivity as reflected in the extent to which subjects learn from reward and punishment feedback as assessed by the probabilistic selection task (PST) and as reflected in the ratio between spontaneous slow-wave theta and fast-wave beta oscillations in the electro-encephalogram. The results indicate that neither of these two methods could be used as a reliable index of individual differences in reward sensitivity. The results of Chapter 3 show that categorization as being reward/punishment sensitive may not solely reflect motivational style, but also the perceptual discriminability or salience of stimulus characters in the PST. The results of Chapter 4 show that high theta/beta EEG ratio is not related to high reward sensitivity, but rather to poor behavioral adaptation in a context with changing reward-punishment contingencies. This may indicate that subjects with a high theta/beta EEG ratio are less able to learn to make optimal decisions based on reward and punishment feedback. In Chapter 5 and 6 it is shown that cortical processes related to the anticipation of reward value and probability are dissociable, statistically, in terms of timing, and neurochemically. More specifically, the results demonstrate that the anticipation of reward involves at least three distinct neural processes specifically related to either the coding of value (reward-related positivity (RRP) and reward P300) or the coding of probability (probability-positivity (PRP)). The value-related processes were found to be under control of dopamine, whereas the probability-related process was found to be under control of noradrenaline. The results described in this thesis may contribute to a better understanding of various mental disorders associated with abnormal valuation of reward or motivational deficits, such as in depression and ADHD.