Figure 7:

Simulation 1b: Separate pathways mediate loss of bursting for reward versus pausing for omission. (a) Empirical results from Matsumoto & Hikosaka (2007), adapted from their Figure 3a with permission from Springer Nature: Nature, copyright 2007, showing flat activity in the LHb following a predicted reward outcome (solid red line). Omitted reward produces phasic increase in activity (dotted blue). (b) Model results showing balanced excitatory inputs to LHbRMTg layer (dash-dot blue line) from VSPatchPosD1 activity (zipper orange) and inhibitory input from PosPV activity (dotted magenta) at the time of predicted reward. While VSPatchPosD1 activity is lower than for PosPV its input to LHbRMTg has a gain factor of 1.7 resulting in an approximate balance. (c) Unopposed input from VSPatchPosD1 activity (zipper orange) at the time of reward omission drives increased LHbRMTg activity (dash-dot blue) and pausing of VTAp dopamine cell firing (solid black).