The flowchart of this framework. (IMAGE)
Caption
(1) The EEG and fNIRS data recordings for all the subjects. (2) The preprocessing steps for cleaning data. (3) The methods for the reconstruction of the signals in the source space. A BEM head model incorporating multiple tissue compartments (scalp, skull, cerebrospinal fluid, and brain) was constructed to simulate the electrical conductivity profiles across heterogeneous head tissues for EEG source localization. Cortical source current density was estimated dipole-wise using sLORETA. Forward modeling integrated Monte Carlo light transport (108 photons) in a 5-layer Colin27 head model with adjoint-derived sensitivity computation, followed by Voronoi-based cortical surface projection for sulcal/gyral-resolved sensitivity mapping. wMNE with spatially adaptive regularization counteracts superficial bias for balanced cortical/deep brain source reconstruction in fNIRS. (4) The EEG and fNIRS source-time series were mapped in the same 3-dimensional space using an atlas-based approach (Desikan–Killiany). (5) FC (Pearson’s correlation) estimates the statistical coupling between each ROI of the reconstructed time series. (6) The topology of brain networks captured by the 2 techniques was compared through graph theoretical approaches.
Credit
Nan Wang, Beijing Tiantan Hospital.
Usage Restrictions
News organizations may use or redistribute this image, with proper attribution, as part of news coverage of this paper only.
License
Licensed content