Since publication, we noticed that the description of our task in the manuscript omitted an aspect of the task design. Specifically, we failed to note that the calorie values that participants viewed during the task contained small deviations from their true value by +/- 15% of total calories for 2/3rd of the trials (1/3rd of the trials showed accurate calorie values) and were rounded to the nearest tens place for all trials. This was done in an attempt to elicit more accurate estimates of calories from participants after the task as, after scanning, participants were told that the calorie labels presented during the scan may have been inaccurate and were then asked to provide their own calorie estimate for each image.

In the presented results, subject-specific calorie estimates were always compared to the true calorie value in all trials. Moreover, as participants were unaware of any deviations in presented calories during scanning, this aspect of the design has no bearing on the analyses we presented in the manuscript that compared trials with calorie information to those without.

The calorie labels are described as “accurate” on pg. 5 under “Stimuli” and in the Fig. 1 caption, but readers should be aware that calories were slightly under- or over-reported for a subset of the stimuli. Both presented and accurate calories are available in our online Open Science Foundation and openneuro.org data repositories.