Figure 2.

Self-control through chaining. (a) A task in which the animal can either take a small reward immediately, or wait and take a larger reward later. Each circle is a state with its value inscribed. We set c=0.2, b₁=1 and b₂=5. In all states, a third action (not represented) causes the animal to leave the task and go back to the initial state (Small reward). (b) Sample simulation of the chaining model on this task. An ‘attempt’ (horizontal axis) is defined as a sequence of actions comprising two successive visits to the initial state. Model parameter values where α_v=0.1, α_w=0.1 and β=2.