Characteristic Analysis of the Resting-State fMRI Global Signal in Schizophrenia

The global signal (GS) represents the broad variations in neural activity throughout the brain. Recent research has identified changes in the GS of functional magnetic resonance imaging (fMRI) in schizophrenia, challenging the traditional view of GS as mere noise that is typically discarded during data preprocessing. However, there has been no comprehensive characteristic analysis of GS in schizophrenia. This study recruited 100 patients with schizophrenia and 113 healthy subjects to perform resting-state fMRI. The averaged gray matter fMRI signal is denoted as GS. The GS topography was constructed by calculating Pearson correlation (i.e. GSCORR) between the GS and time series of each gray matter voxel. Furthermore, the relevance between GS topography and the function network features was constructed according to graph theory. Finally, we implemented an integrated analytical framework combining independent component analysis and multiple linear regression to quantify the contributions of resting-state brain networks to the spatiotemporal characteristics of the GS. The GSCORR of schizophrenia decreased in the bilateral insula and exhibited a significant negative association with disease duration. The increased GSCORR in the thalamus and default mode network (DMN) showed a positive correlation with the scale scores. We found that global properties are represented by GS topography in schizophrenia. In addition, the linear characterization results of GS showed that the component contribution decreased at the primary sensory network but increased at the high-order associated network. Our results further demonstrated that GS contains brain features associated with schizophrenia, which would help to understand the neural mechanisms underlying the psychopathological symptoms of schizophrenia.