Four hypotheses about state-space and action representations.

(a) The state-space constitutes a single stage (outcomes such as food pellets are not shown here as states), and ‘L’ and ‘R’ are the only possible actions that the subject considers taking. (b) The state-space matches the correct state-space of the task, and the actions are ‘L’ and ‘R’. (c) The state-space only consists of state S0, and actions are ‘L → R’ and ‘R → L’. Note that single actions ‘R’ and ‘L’ are not included. (d) The state-space represents the two stages of the task and actions include both single actions and action sequences.