test and verify such a scenario, an alternative possibility is that gamma-50 reflects different settings in which reward-related processing is engaged. There is a growing literature documenting representations of reward not being activated only by reward receipt itself, but also during movement preparation (German and Fields, 2007; Jin and Costa, 2009; Lauwereyns et al., 2002; Taha et al., 2007), decision-making (van der Meer and Redish, 2009), and offline processing (Lansink et al., 2008). Reinforcement learning models require explicit links between actions and action values (Sutton and Barto, 1998), which need to be updated following rewards, as well as retrieved during action selection. The results reported here are consistent with the idea that ventral striatal gamma oscillations play a role in both the formation and recall of associations relevant to motivated behavior. The observed separation between gamma-50 and gamma-80, their complex behavioral correlates, and their relationship with spiking activity likely reflect functionally important aspects of ventral striatal processing, and provide entry points for further investigation.
