MEG Coherence and DTI Connectivity in mTLE

Abstract

Magnetoencephalography (MEG) is a noninvasive imaging method for localization of focal epileptiform activity in patients with epilepsy. Diffusion tensor imaging (DTI) is a noninvasive imaging method for measuring the diffusion properties of the underlying white matter tracts through which epileptiform activity is propagated. This study investigates the relationship between the cerebral functional abnormalities quantified by MEG coherence and structural abnormalities quantified by DTI in mesial temporal lobe epilepsy (mTLE). Resting state MEG data was analyzed using MEG coherence source imaging (MEG-CSI) method to determine the coherence in 54 anatomical sites in 17 adult mTLE patients with surgical resection and Engel class I outcome, and 17 age- and gender- matched controls. DTI tractography identified the fiber tracts passing through these same anatomical sites of the same subjects. Then, DTI nodal degree and laterality index were calculated and compared with the corresponding MEG coherence and laterality index. MEG coherence laterality, after Bonferroni adjustment, showed significant differences for right versus left mTLE in insular cortex and both lateral orbitofrontal and superior temporal gyri (p < 0.017). Likewise, DTI nodal degree laterality, after Bonferroni adjustment, showed significant differences for right versus left mTLE in gyrus rectus, insular cortex, precuneus and superior temporal gyrus (p < 0.017). In insular cortex, MEG coherence laterality correlated with DTI nodal degree laterality ($R^{2} = 0.46; p = 0.003)$ in the cases of mTLE. None of these anatomical sites showed statistically significant differences in coherence laterality between right and left sides of the controls. Coherence laterality was in agreement with the declared side of epileptogenicity in insular cortex (in 82 % of patients) and both lateral orbitofrontal (88 %) and superior temporal gyri (88 %). Nodal degree laterality was also in agreement with the declared side of epileptogenicity in gyrus rectus (in 88 % of patients), insular cortex (71 %), precuneus (82 %) and superior temporal gyrus (94 %). Combining all significant laterality indices improved the lateralization accuracy to 94 % and 100 % for the coherence and nodal degree laterality indices, respectively. The associated variations in diffusion properties of fiber tracts quantified by DTI and coherence measures quantified by MEG with respect to epileptogenicity possibly reflect the chronic microstructural cerebral changes associated with functional interictal activity. The proposed methodology for using MEG and DTI to investigate diffusion abnormalities related to focal epileptogenicity and propagation may provide a further means of noninvasive lateralization.

Appendix

The following steps are taken for gray matter modeling:

• 1. Create the AC-PC (anterior commissure- posterior commissure) coordinate system using the graphical interface to identify structural landmarks (Fig. 14).

  Fig. 14

  

  AC-PC coregistration (a) and graphical tools in MEG Tools for MEG-MRI coregistration (b) and (c). In a, the anterior commissure and, in b, the posterior commissure are identified

• 2. Identify the outline of the cortical surface on five MRI slices allowing the user to compensate for defects (Fig. 15). The slices are interpolated to a total number of 15 cortical slice boundaries.

  Fig. 15

  

  The location of the cortical boundary is manually identified in five slices

• 3. Using linear and locally nonlinear transforms a smooth cortical surface model is fit to the 15 cortical boundaries, including five user-drawn boundaries in Fig. 15 (Fig. 16).

  Fig. 16

  

  A smooth cortical surface model of the subject (in gray scale) is constructed to minimize the surface distance with 15 cortical boundaries. The cortical boundaries are depicted in blue, including five boundaries drawn by the user. The outer border of the gray matter in the cortical source model slices are depicted by red lines (Color figure online)

• 4. The cortical gray matter is identified and a 4000 source location cortical model constructed for MEG imaging. The cortical surface model of the subject is adjusted to match the outer boundary of the cortical gray matter.

• 5. The cortical surface is transformed to AC-PC coordinates. A combination of linear warps and shears is applied to the subject’s cortical surface to achieve the best match to a cortical surface model of the MNI305 brain. The shears align the anterior and posterior poles of the subject cortical surface with the MNI surface model. A closest neighbor algorithm is used to identify the corresponding subject and MNI surface points which are used in all transform calculations. These transforms operate on the volume within the cortical surface as well as the surface itself.

• 6. Within a sequence of three to five overlapping Gaussian windows, second order transforms of included brain volume are calculated along the inferior-superior axis (Z) first, left–right axis (X) second and posterior-anterior (Y) axis last (Fig. 17). These locally nonlinear brain volume transforms further optimize the match of the subject cortical surface to the MNI surface model within each window (Fig. 18), using the following equation:

  Fig. 17

  

  Five overlapped window functions are shown along the posterior-anterior axis of the cortical surface. The maximum amplitude of each window function is 1. Adjacent window functions overlap and the sum of their amplitudes is one. Thus, the full transform throughout the cortical volume is a continuous mixture of windowed transforms. The same windowing technique is applied along the inferior-superior axis and left–right axis. Furthermore, these one dimensional windows can be altered to be two- or three-dimensional windows to achieve more focal sensitivity to mismatch and to accommodate internal structural matching (Color figure online)

  Fig. 18

  

  Errors in the transformed cortical surface are shown after the initial linear transformation and after the final nonlinear transformations. The magnitude of the error is displayed in color on the MNI305 cortical surface, with red corresponding to maximum error (Color figure online)

$$\left[ \begin{aligned} X_{MNI} \hfill \\ Y_{MNI} \hfill \\ Z_{MNI} \hfill \\ \end{aligned} \right] = \left( {\begin{array}{*{20}c} {a_{xc} } & \ldots & {a_{xyz} } \\ \vdots & \ddots & \vdots \\ {a_{zc} } & \cdots & {a_{zyz} } \\ \end{array} } \right)\left[ \begin{array}{ll} C_{X} \hfill \\ C_{Y} \hfill \\ C_{Z} \hfill \\ X \hfill \\ Y \hfill \\ Z \hfill \\ X^{2} \hfill \\ Y^{2} \hfill \\ Z^{2} \hfill \\ XY \hfill \\ XZ \hfill \\ YZ \hfill \\ \end{array} \right] \, with \, {\rm A} = \left( {\begin{array}{*{20}c} {a_{xc} } & \ldots & {a_{xyz} } \\ \vdots & \ddots & \vdots \\ {a_{zc} } & \cdots & {a_{zyz} } \\ \end{array} } \right)$$

(5)

where A is transform matrix, XYZ, and ${\text{X}}_{\text{MNI}} {\text{Y}}_{\text{MNI}} {\text{Z}}_{\text{MNI}}$ are the native and MNI coordinate systems, respectively. The MNI coordinate system specifies the location of the brain structure within an AC-PC coordinate system (Fig. 14). Generalized Gaussian window functions are used to eliminate transform discontinuities. The inverse transform is generated to convert MRI to MNI305 coordinates. The algorithm generates a set of sequentially applied transforms that are applied to the original MRI pixel coordinates of the MEG imaging results.