Abstract
The weighted Phase Lag Index (wPLI) and the weighted Symbolic Mutual Information (wSMI) represent two robust and widely used methods for MEG/EEG functional connectivity estimation. Interestingly, both methods have been shown to detect relative alterations of brain functional connectivity in conditions associated with changes in the level of consciousness, such as following severe brain injury or under anaesthesia. Despite these promising findings, it was unclear whether wPLI and wSMI may account for distinct or similar types of functional interactions. Using simulated high-density (hd-)EEG data, we demonstrate that, while wPLI has high sensitivity for couplings presenting a mixture of linear and nonlinear interdependencies, only wSMI can detect purely nonlinear interaction dynamics. Moreover, we evaluated the potential impact of these differences on real experimental data by computing wPLI and wSMI connectivity in hd-EEG recordings of 12 healthy adults during wakefulness and deep (N3-)sleep, characterised by different levels of consciousness. In line with the simulation-based findings, this analysis revealed that both methods have different sensitivity for changes in brain connectivity across the two vigilance states. Our results indicate that the conjoint use of wPLI and wSMI may represent a powerful tool to study the functional bases of consciousness in physiological and pathological conditions.
Highlights
IntroductionFunctional connectivity (FC) metrics identify statistical (undirected) associations among spatially distinct brain areas
Functional connectivity (FC) metrics identify statistical associations among spatially distinct brain areas
For each pair of source locations (LIPL-right middle frontal gyrus (RMFG) and RIPL-RMFG) and each type of simulated source coupling dynamics we modelled 100 different signal-to-noise ratios (SNR; from 0.01 to 1, with steps of 0.01), which describe the weighting of simulated source signals with respect to simulated background activity
Summary
Functional connectivity (FC) metrics identify statistical (undirected) associations among spatially distinct brain areas. The weighted Phase Lag Index (wPLI1) and the weighted Symbolic Mutual Information (wSMI4), represent examples of spectral (wPLI) and information-theoretic (wSMI) connectivity estimation methods that are increasingly applied to both EEG and MEG data[5,6,7,8,9,10,11,12,13,14,15,16,17,18,19] These two connectivity metrics are modified versions of pre-existing methods (PLI1,20; SMI4) that minimise the contribution of ‘(almost-)zero-lag’ interactions, potentially determined by volume conduction. Www.nature.com/scientificreports of underlying brain sources, while excluding apparent zero-lag connectivity driven by a mixture of real and spurious relationships[26,27] Both wPLI and wSMI have been applied to explore brain functional dynamics associated with different behavioural states[6,12] or potential network-level alterations in pathological conditions (e.g., Alzheimer’s disease[13], multiple sclerosis[14], schizophrenia[15] and social anxiety disorder[16]). Here we asked whether the two connectivity metrics provide overlapping or complementary information about changes in brain functional dynamics across the two vigilance states
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call