Figure 2.

Results of Pre-post Assessments

All plots show Δ, the change in performance (post − pre, baseline corrected) in the trained and control groups. Error bars represent standard errors. Asterisks indicate significance of ANCOVA analyses as described in Results: ^∗p < 0.05; ^∗∗p < 0.01; (^∗)p < 0.10.

(A) Cognitive control. Graph shows proportion of correct anticipatory looks in the preswitch (initial rule learning) and postswitch (unlearning one rule and learning another) phases.

(B) Gap overlap. “Average RT” is the average of the three conditions that we administered (gap, baseline, and overlap). “Facilitation” shows the facilitation effect, and “disengagement” shows disengagement latencies. Because the valence of the observed changes in the task was negative, −Δ values are presented for ease of comparison.

(C) Sustained attention. “Mixed dynamic/static” shows the results of experiment 1, which measured looking behavior toward a mixture of dynamic and static stimuli. “Interesting static” and “boring static” show the results of experiment 2, which measured looking behavior toward “interesting” and “boring” static images.

(D) Working memory. Graph shows median delay length for trials followed by a correct response.

(E) Structured free play. “Looks to object” shows number of separate looks to the target objects. “Shifts from object to person” shows number of attention shifts from looking at the objects to looking at either the experimenter or caregiver. “Average duration of looks to object” shows the average duration of looks toward the target objects.