Distinct effects of global signal regression on brain activity during propofol and sevoflurane anesthesia.

Abstract

Global signal regression (GSR) is widely used in functional magnetic resonance imaging (fMRI) analysis, yet its effects on anesthetic-related brain activity are not well understood. Using fMRI data from patients under general anesthesia, we analyzed temporal variability indices, amplitude of low-frequency fluctuations, functional connectivity, and graph theoretical measures with and without GSR. Here we show that GSR differentially affects brain activity patterns during propofol- and sevoflurane-induced unconsciousness. While temporal variability indices decreased similarly between conscious and unconscious states regardless of GSR, functional connectivity analyses revealed anesthetic-specific effects: GSR altered specific network connections under propofol but broadly reduced connectivity differences under sevoflurane. Network topology analyses demonstrated that GSR minimally affected propofol-induced changes in graph theoretical measures but significantly diminished sevoflurane-related network alterations. These findings reveal that GSR's impact on functional brain organization is anesthetic-specific, with sevoflurane-induced changes being particularly sensitive to global signal removal. Our results suggest that GSR should be applied cautiously when comparing different anesthetic agents and highlight the importance of considering drug-specific effects when analyzing consciousness-related brain activity.