Figure 8:

Simulation 1c: Asymmetric dopamine signaling for late-versus-early reward. (a) Empirical results adapted from Hollerman and Schultz (1998), Figure 6b with permission from Springer Nature: Nature Neuroscience, copyright 1998, showing an asymmetric pattern of firing for late (thin arrow) versus early (thick arrow) reward delivery. (b,c) Simulation results for late-versus-early reward respectively, capturing the empirical results. (d) Focus on the USTime_In input layer, representing the OFC bridging between CS and US, with a temporally-evolving, US-specific pattern that drives the VS patch expectations of US timing. When the US arrives early, it resets this US timing representation, thereby preventing VS patch firing.