In the article “Task Encoding across the Multiple Demand Cortex Is Consistent with a Frontoparietal and Cingulo-Opercular Dual Networks Distinction,” by Ben M. Crittenden, Daniel J. Mitchell, and John Duncan, which appeared on pages 6147–6155 of the June 8, 2016 issue, there was an error in our description of the method of data normalization before pattern classification. We used the default method employed by the software toolbox (the decoding toolbox), however due to a misunderstanding of the code, we reported the normalization incorrectly. Specifically, where we state that “The resulting β values from each task produced by the subject-wise fixed effects analyses were Z-scored across all voxels within each ROI,” this should have read, “The resulting β values from each task produced by the subject-wise fixed effects analyses were Z-scored across tasks for each voxel within the ROI.” The results remain unchanged.