Figure 1.
DTI and network construction based on the Automated Anatomical Labelling (AAL) parcellation [28].
Figure 2.
DBS electrode artefact from lead.
The artefact from the external lead of the DBS electrodes is visible in the form of dropout in left hemisphere in A) the b0 weighted image and B) the diffusion tensor image.
Figure 3.
Anatomical connectivity networks derived from DTI data.
The pre- and post-operative structural networks (left and right columns, respectively) are shown superimposed on a rendered brain (A/B) and as connectivity matrices, Cpre and Cpost (C/D). In both representations, the full 90-node networks are highlighted in green, while the left and right hemispheres are highlighted in blue and red, correspondingly (In Table 2 we report the indexing of brain areas). In the matrix representations, transcallosal connectivity is shown in the other diagonals. In C and D, the red arrows point to the pre- and post-operative 45×45 right hemisphere connectivity matrices.
Table 1.
Graph properties of the preoperative, postoperative and control structural networks (right hemisphere only).
Figure 4.
Nodal efficiency changes between pre- and post-DBS structural networks.
The AAL regions with more than 20% difference in nodal efficiency between pre- and postoperative measures are plotted on three-dimensional renderings of the human brain in MNI space seen from above (top) and from the side (bottom). The size and colour of the circles indicate the magnitude (size) of the increases (blue) and decrease (red) in nodal efficiency after DBS. The number inside each circle indicates the AAL ordering index reported in Table 2.
Table 2.
Listing the 45Figure 3.
Figure 5.
Exploring the impact of DBS-induced structural changes on resting-state functional connectivity.
A) Schematic overview of how the simulated FC matrices are obtained from the SC matrices using the dynamic mean field (DMF) model. The simulated FC matrices are subsequently compared to the empirical resting state FC matrix. B) Solid lines indicate the fitting of simulated functional connectivity (FC) matrices obtained with the pre-DBS (black), post-DBS (red), and healthy controls (blue) structural connectivity matrices with the empirical healthy FC, as a function of the global coupling weight (G). Vertical dashed lines indicate the corresponding bifurcation points, above which the dynamics becomes unstable. We observe that the bifurcation point of the post-DBS FC is shifted from the pre-DBS FC bifurcation point towards the healthy bifurcation point. This means that, before DBS, the structural connectivity is weaker and therefore stronger couplings are required to reach an optimal fitting with empirical FC The shift of the post-DBS bifurcation point towards the healthy bifurcation point indicates recovery of the structural connectivity with DBS.
Table 3.
Graph properties of simulated functional networks derived from the dynamic mean field model with the 3 structural connectomes from the preoperative PD, postoperative PD and healthy controls.