The study investigates the validity of anticorrelations observed in resting-state functional magnetic resonance imaging (fMRI) data, which have been proposed to exist between the default mode network and the task-positive network. The authors compared two preprocessing methods: global signal regression and a component-based noise reduction method (CompCor). They found that robust anticorrelations between the default mode network and the task-positive network were observed under both methods, suggesting that these anticorrelations are not artifacts introduced by global signal regression. The specificity of anticorrelations was similar between the two methods, but the CompCor method showed higher specificity and sensitivity for positive correlations. The results indicate that anticorrelations in resting-state connectivity are likely biologically significant and that the CompCor method can effectively remove spurious noise, making it a suitable alternative to global signal regression.The study investigates the validity of anticorrelations observed in resting-state functional magnetic resonance imaging (fMRI) data, which have been proposed to exist between the default mode network and the task-positive network. The authors compared two preprocessing methods: global signal regression and a component-based noise reduction method (CompCor). They found that robust anticorrelations between the default mode network and the task-positive network were observed under both methods, suggesting that these anticorrelations are not artifacts introduced by global signal regression. The specificity of anticorrelations was similar between the two methods, but the CompCor method showed higher specificity and sensitivity for positive correlations. The results indicate that anticorrelations in resting-state connectivity are likely biologically significant and that the CompCor method can effectively remove spurious noise, making it a suitable alternative to global signal regression.