Skip to main content
. 2020 Jan 22;105(2):370–384.e8. doi: 10.1016/j.neuron.2019.10.030

Figure 2.

Figure 2

Animals Used Contextual Factors to Decide Whether and When to Act

(A–E) The effect of present and past contextual factors on actTime (left y axis). The corresponding reward probability at response is displayed on the right y axis. actTime was longer in long (7 s) and medium (5 s) compared with short (3 s) ITI blocks (A); in fast compared with slow trials (B); when they were offered a large compared with a medium reward and when offered a small compared with a medium reward, giving rise to a U-shaped effect of reward magnitude on actTime (C); when a previous trial was not rewarded (D); and when they had already delayed actTime on the past trial (E). For illustrative purposes, actTime in the immediate past trial is binned into three groups. Each color represents one animal, and each ring is one testing session. The gray columns illustrate the group mean across all observations. Multilevel ANOVA followed by Tukey’s HSD.

(F) The log of the odds of responding during the trial for different levels of present and past contextual factors.

p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

See also Figure S1.