The coding of unexpected events and prediction errors is essential for efficient behavior as well as necessary for trial-and-error based learning to occur. Given this, understanding where in the brain the coding takes place is highly significant. The midbrain dopamine system projecting to the striatum has been implicated in processing prediction errors, however it remains unclear if this system codes reward prediction errors specifically, or more generally, prediction errors of behaviorally relevant events that interrupt and require a response. The main goal of this study is to use fMRI to investigate processing of unexpected events (including prediction errors) in the human brain and the influence that the behavioral relevance of the stimuli has on the processing. We hypothesize that striatal activity will be modulated by unexpected behaviorally relevant stimuli and prediction errors, irrespective of reward value; striatal activity will not be modulated when the stimuli are behaviorally irrelevant. To test this hypothesis, the following aims will be addressed.
AIM #1 : To investigate how unexpected rewarding stimuli (money) modulate human brain activity when the stimuli are behaviorally relevant and when they are behaviorally irrelevant;
AIM #2 : To investigate how human brain activity is modulated by temporal prediction errors of neutral visual stimuli, when the stimuli are behaviorally relevant and when they are behaviorally irrelevant. In a preliminary study, activity in the striatum increased following unexpected (infrequent), non-rewarding visual stimuli, only when the stimuli were behaviorally relevant, i.e. potentially required a response. ? ?
| Zink, Caroline F; Pagnoni, Giuseppe; Chappelow, Jonathan et al. (2006) Human striatal activation reflects degree of stimulus saliency. Neuroimage 29:977-83 |
| Zink, Caroline F; Pagnoni, Giuseppe; Martin-Skurski, Megan E et al. (2004) Human striatal responses to monetary reward depend on saliency. Neuron 42:509-17 |
