Abstract :
[en] A major topic in fMRI analysis is how to address the global signal (GS). The GS refers to the mean time course computed over all voxels within the brain. As such, it has been often treated as noise during preprocessing. Here, we would like to advocate for the alternative view, i.e. that the GS is an informative source for cognitive function and behavior. To do so, we bring together our diverse expertise to raise a comprehensive update on the GS, by considering: 1) its basic properties, 2) its mathematical influence on functional dynamics, 3) its manifestation across lifespan, cognition and behavior, and 4) its assistance in better interpreting ongoing mental states. We believe that such discussion will be beneficial for network neuroscience investigations, as it aims to highlight the need to carefully consider whether or not to model the GS as typical noise source.
Commentary :
PRESENTATIONS
Thomas Liu
Departments of Radiology, Psychiatry, and Bioengineering
University of California San Diego, La Jolla (CA)
The Global Signal and Its Role in Functional MRI
The global signal is widely used as a regressor or normalization factor for removing the effects of global variations in the analysis of functional magnetic resonance imaging (fMRI) experiments. However, there continues to be considerable controversy regarding its use and interpretation. In this talk, I will review the basic properties of the global signal and describe the various ways that it has been used for the analysis of both task-related and resting-state fMRI data. I will also discuss the sources of information that are embedded in the global signal and touch upon emerging views of its role in the analysis of fMRI studies.
Dimitri Van De Ville
École Polytechnique Fédérale de Lausanne, Neuro-X Institute, Geneva
University of Geneva, Department of Radiology and Medical Informatics, Geneva
Global Signal and Its Impact on Dynamic Functional Connectivity
Global signal regression (GSR) is often considered as a preprocessing step for resting-state fMRI data. The many different analysis pipelines for dynamic functional connectivity (dFC) treat the global signal implicitly through centering, normalization, clustering, and other operations. Therefore, in some cases, the global signal can be separated by the analysis method without the need for GSR as a preprocessing step. In addition, dFC allows to extract properties of the spatiotemporal organization of the global signal that can provide insights into its role and function.
Lucina Q. Uddin
Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles (CA)
Functional significance of the global signal across the lifespan
Spatiotemporal patterns resembling global signal (GS) topography were shown to explain > 20% of the variance in intrinsic BOLD timeseries (Bolt et al., 2022). Such GS topography was also mediated by individual differences in positive/ negative life outcomes and psychological function (Li et al., 2019). More recently, we found systematic age-associations, where subcortical vs. cortical contributions to the GS topography differed across the lifespan (Nomi et al., 2023). These results suggest that the GS contains rich information related to trait-level cognition, highlighting the need to carefully consider whether or not to remove the GS during preprocessing.
Sepehr Mortaheb
GIGA-CRC In Vivo Imaging, University of Liège, Liège, Belgium
Role of the Global Signal in Mental State Characterization
The fMRI global signal (GS) as the average BOLD signal across the brain, is a controversial aspect of the brain’s functional data analysis. While some argue that the GS reflects non-neural physiological noise, others contend it contains valuable information about the brain’s functional organization. We recently showed that the GS amplitude varies across different mental states, from mind blanking (Mortaheb et al 2022; Boulakis et al, 2023) to psychedelic experiences (Mortaheb et al, 2023). Here, I will present how the GS amplitude complements the dynamic functional connectivity analysis to better interpret the neural substrate of different mental states.
