The authors regret the following error reported in our methodology: it was not accurate to state that the width of the psychophysical gaze curve was calculated using the eye contract signal strength at the 5% and 95% eye-contact endorsement values as stated on page 5 and Figure 2c of the article. Rather, our specific psychometric function was constrained to asymptotes at 0.02 and 0.98 (range = 0.96). As a result, the “5% from the asymptotes” for a scale of 0.96 becomes 4.8% from the asymptotes in actual eye-contact endorsement value. Consequently, the eye-contact endorsement values used to derive the width of the curve become 0.02 + 0.048 = 0.068 (6.8%) and 0.98 − 0.048 = 0.932 (93.2%) instead of 0.05 (5%) and 0.95 (95%).

This does not change the results or findings, but the authors would still like the publication to accurately reflect our methodology. The authors apologize for any inconvenience caused.

The editors accept this correction and are grateful to the authors for reporting it.