Arm maze simulation

This model explores how the Rubicon model exhibits goal-driven decision-making in the case of a simulated rat in a multi-arm maze, where each arm has different cost vs. benefit tradeoffs that must be learned and then used to make better decisions.

Friedman et al., 2015 shows that PL projections to dorsomedial striatum (ALM / dlPFC motor planning area) activate inhibitory interneurons in striosomes, which then inhibit striosomes, during high-conflict cases where cost-benefit ratio is high. Inhibiting this PL projection led to an increase in high-reward choices specifically in the high-cost, high-reward case. Activating PL had opposite effects. Activating ACC lead to increase in higher-reward preference across the board.