Correction to Molecular Psychiatry https://doi.org/10.1038/mp.2014.102, published online 09 September 2014

We recently noticed a software error affecting the conversion of the DICOM files generated by the PET scanner to the NIfTI files used in our analysis pipeline. Correcting this error did not affect the results or conclusions described in the paper’s Abstract or the primary results of the paper regarding voxel-wise correlation patterns between striatal dopamine D2-like binding potential (D2BP) with body mass index (BMI) and opportunistic eating behavior (see Supplement for updated voxel-wise correlation maps). However, as we discussed in the paper, averaging over regions of interest (ROI) can obscure the positive and negative correlation patterns between striatal D2BP and BMI and the corrected ROI analyses no longer indicated significant differences in D2BP between participants with and without obesity. Specifically, mean ± 95% CI for D2BP was 28.0 ± 1.5 and 28.3 ± 1.1 (p = 0.73) in caudate, 30.6 ± 1.7 and 31.0 ± 1.3 (p = 0.69) in putamen, and 19.6 ± 1.8 and 18.8 ± 1.1 (p = 0.42) in nucleus accumbens for participants with and without obesity, respectively. The corrected Pearson correlation coefficients between D2BP in striatal ROIs with BMI were also not significant, with r = 0.018 (p = 0.91) in caudate, r = 0.089 (p = 0.57) in putamen, and r = 0.14 (p = 0.37) in nucleus accumbens. We regret this error and apologize for any inconvenience it has caused.