Journal Article

On the interplay between state-dependent reconfigurations of global signal correlation and BOLD fluctuations: An fMRI study

Ricca, Valdo, Damiani, Stefano, Scalabrini, Andrea, La-Torraca-Vittori, Paolo, Politi, Pierluigi, Tarchi, Livio, Tosi, Eleonora, Fusar-Poli, Paolo

2024

Overview

•From rest to task, GSCORR reduces in sensory but increases in associative areas.•Reallocation of functional resources from sensory to associative regions at task.•Two distinct clusters of areas are recruited during GO and STOP trials.•GSCORR correlates with deactivation rather than activation.•GSCORR, activation and deactivation reflect distinct neurofunctional processes. The dynamics of global, state-dependent reconfigurations in brain connectivity are yet unclear. We aimed at assessing reconfigurations of the global signal correlation coefficient (GSCORR), a measure of the connectivity between each voxel timeseries and the global signal, from resting-state to a stop-signal task. The secondary aim was to assess the relationship between GSCORR and blood-oxygen-level-dependent (BOLD) activations or deactivation across three different trial-conditions (GO, STOP-correct, and STOP-incorrect). As primary analysis we computed whole-brain, voxel-wise GSCORR during resting-state (GSCORR-rest) and stop-signal task (GSCORR-task) in 107 healthy subjects aged 21–50, deriving GSCORR-shift as GSCORR-task minus GSCORR-rest. GSCORR-tr and trGSCORR-shift were also computed on the task residual time series to quantify the impact of the task-related activity during the trials. To test the secondary aim, brain regions were firstly divided in one cluster showing significant task-related activation and one showing significant deactivation across the three trial conditions. Then, correlations between GSCORR-rest/task/shift and activation/deactivation in the two clusters were computed. As sensitivity analysis, GSCORR-shift was computed on the same sample after performing a global signal regression and GSCORR-rest/task/shift were correlated with the task performance. Sensory and temporo-parietal regions exhibited a negative GSCORR-shift. Conversely, associative regions (ie. left lingual gyrus, bilateral dorsal posterior cingulate gyrus, cerebellum areas, thalamus, posterolateral parietal cortex) displayed a positive GSCORR-shift (FDR-corrected p < 0.05). GSCORR-shift showed similar patterns to trGSCORR-shift (magnitude increased) and after global signal regression (magnitude decreased). Concerning BOLD changes, Brodmann area 6 and inferior parietal lobule showed activation, while posterior parietal lobule, cuneus, precuneus, middle frontal gyrus showed deactivation (FDR-corrected p < 0.05). No correlations were found between GSCORR-rest/task/shift and beta-coefficients in the activation cluster, although negative correlations were observed between GSCORR-task and GO/STOP-correct deactivation (Pearson rho=-0.299/-0.273; Bonferroni-p < 0.05). Weak associations between GSCORR and task performance were observed (uncorrected p < 0.05). GSCORR state-dependent reconfiguration indicates a reallocation of functional resources to associative areas during stop-signal task. GSCORR, activation and deactivation may represent distinct proxies of brain states with specific neurofunctional relevance.

• Publisher: Elsevier Inc,Elsevier Limited,Elsevier
• Subject: Anxiety disorders / Bipolar disorder / Blood levels / Bold activation / Brain mapping / Brodmann's area / Cerebellum / Cortex (parietal) / Deactivation / Frontal gyrus / Functional connectivity / Functional magnetic resonance imaging / Global signal / Medical imaging / Mental depression / Negative BOLD signals / Neural networks / Neuroimaging / Resting state / Schizophrenia / Sensitivity analysis / Somatosensory cortex / Stop-signal task / Time series / Topography