An additional consideration that almost certainly contributed to differences in results is the stage of reward processing that was analyzed. All of these fMRI tasks included three basic stages: cue presentation, anticipation of reward following behavioral response, and feedback. Of the studies reviewed here, three studies examined anticipation of reward (Bjork et al., 2004; Galván et al., 2006; Eshel et al., 2007), three studies analyzed responses to feedback (Bjork et al., 2004; Ernst et al., 2005; van Leijenhorst et al., 2009) and one study did not distinguish between stages and instead analyzed the entire trial (May et al., 2004). The difficulty in analyzing these different stages of reward processing is that temporally proximal events (e.g., the cue and anticipation phase) are difficult to parse out in fMRI analyses. In practice, this means that while only one phase was of interest, MR signal from the other stages may have bled into activation. In other words, while researchers may have intended to examine one aspect of the task, they may have been measuring (and reporting) another aspect of the task. Without the
