This study investigates whether anticorrelations observed in resting-state functional magnetic resonance imaging (fMRI) are biologically valid or artifacts introduced by global signal regression. The researchers compared two preprocessing methods: CompCor, which removes noise from white matter and cerebrospinal fluid regions, and global signal regression. They found robust anticorrelations between the default mode network and task-positive network regions using both methods, suggesting that these anticorrelations are not artifacts but may have biological origins. The CompCor method provided higher specificity for positive correlations compared to global signal regression. The study concludes that anticorrelations in resting-state connectivity are not solely due to global signal regression and may reflect genuine neural interactions. The findings support the use of CompCor for more accurate analysis of resting-state connectivity.This study investigates whether anticorrelations observed in resting-state functional magnetic resonance imaging (fMRI) are biologically valid or artifacts introduced by global signal regression. The researchers compared two preprocessing methods: CompCor, which removes noise from white matter and cerebrospinal fluid regions, and global signal regression. They found robust anticorrelations between the default mode network and task-positive network regions using both methods, suggesting that these anticorrelations are not artifacts but may have biological origins. The CompCor method provided higher specificity for positive correlations compared to global signal regression. The study concludes that anticorrelations in resting-state connectivity are not solely due to global signal regression and may reflect genuine neural interactions. The findings support the use of CompCor for more accurate analysis of resting-state connectivity.