Figure 2

Pupil dilation (Experiment 1) and Brain activations (Experiment 2) during the Scooping and Object Out phases. At Scooping, one of two objects has entered the cup either in a visible (Known) or invisible (Unknown) manner. At Object Out, the object behind the occluder exits. Depending on the condition, the exit may be informative to deduce the identity of the object inside the cup (INF: inference needed; LOI: no inference can be drawn: NOV and NON: no inference needed). (A–D) Temporal course of the mean pupil dilation (SEM) from baseline. Grey horizontal lines indicate areas of effect identified by permutation tests, (B–D) additionally Bonferroni-corrected. (A) Unknown (purple) and Known (dark blue) conditions. Pupil size reduced in the Known compared to the Unknown condition, indicating sustained cognitive effort due to possible logical processing in the latter. (B) INF (red) and NOV (blue stone) conditions; (C) INF and LOI (light red) conditions; and (D) LOI and NON (light blue) conditions. Pupil dilation was higher when a logical inference was required to identify the object in the cup, or when an unknown representation had to be maintained, compared to the conditions requiring no inference (B, D respectively). Likewise, a logical inference elicited a higher pupil dilation compared to LOI, another condition with logical valence (C). (E–H) Whole brain views (lateral left and right, posterior and medial) of the main brain areas responding to the (E) Unknown (orange/yellow) and Known (blue/green) conditions (T > 3.47); to the INF condition (orange/yellow) contrasted to (F) the NOV (blue/green) and (G) the LOI (blue/green) conditions; and to the LOI contrasted to the NON condition (H) (T > 3.09). See Supplementary Table S1−S6 for other active regions.