Skip to main content
. 2020 Mar 9;10:4287. doi: 10.1038/s41598-020-61257-y

Figure 1.

Figure 1

(a) Schematic illustration of a low and a high reward trial in the task. After presentation of a cue, a rapid sequence of target episodes was presented rapidly during the foraging patch. (b) Computer simulation of the two signal sources in the temporal difference model investigated in the paper (see section 2.2 of the Methods for details). The figure shows time-averaged signals generated by the algorithm after training. The prediction error signal δ, in green in the figure, represents changes in expected rewards, while r is the amount of reward effectively available in the foraging patch (in red). Smoothed over time, r gives the reward rate of the foraging patch. According to the model, the green curve corresponds to the incentive motivation signal, and the red curve to reward rates/perceived opportunities.