A novel approach for global noise reduction in resting-state fMRI: APPLECOR

► We introduce a new approach to data-driven physiological noise correction for fMRI. ► The correction increases the consistency of default mode network correlations. ► Resting state consistency was improved both within and across subjects. ► Greater consistency was obtained with the new approach than with other corrections. ► Combining the new approach with a model-driven correction yielded the best results.