Alterations in brain structure associated with trigeminal nerve anatomy in episodic migraine

Abstract

The pathophysiology of migraine remains to be elucidated. We have recently shown that interictal migraineurs exhibit reduced fractional anisotropy (FA) in the root entry zone of the trigeminal nerve when compared to controls, but it is not known if this altered nerve anatomy is associated with changes within the brainstem or higher cortical brain regions. Diffusion tensor imaging of the brain was used to calculate regional measures of structure, including mean diffusivity (MD), axial diffusivity (AX) and radial diffusivity (RD) in addition to voxel-based morphometry of T1-weighted anatomical images. Linear relationships between trigeminal nerve anatomy (FA) and MD throughout the brainstem and/or higher cortical regions were determined in both controls (n = 31, brainstem; n = 38, wholebrain) and interictal migraineurs (n = 32, brainstem; n = 38, wholebrain). Additionally, within the same brain areas, relationships of AX and RD with nerve FA were determined. We found that in both interictal migraine and control participants, decreasing trigeminal nerve FA was associated with significantly increased MD in brainstem regions including the spinal trigeminal nucleus and midbrain periaqueductal gray matter (PAG), and in higher brain regions such as the hypothalamus, insula, posterior cingulate, primary somatosensory and primary visual (V1) cortices. Whereas, both control and migraineur groups individually displayed significant inverse correlations between nerve FA and MD, in migraineurs this pattern was disrupted in the areas of the PAG and V1, with only the control group displaying a significant linear relationship (PAG controls r = -0.58, p = 0.003; migraineurs r = -0.25, p = 0.17 and V1 controls r = -0.52, p = 0.002; migraineurs r = -0.10, p = 0.55). Contrastingly, we found no gray matter volume changes in brainstem or wholebrain areas. These data show that overall, trigeminal nerve anatomy is significantly related to regional brain structure in both controls and migraineurs. Importantly, the PAG showed a disruption of this relationship in migraineurs suggesting that the anatomy and possibly the function of the PAG is uniquely altered in episodic migraine, which may contribute to altered orofacial pain processing in migraine.

Keywords: MRI; PAG; brain imaging; diffusion tensor imaging; fractional anisotropy; mean diffusivity.

Figure 1

Axial T1-weighted anatomical image and corresponding diffusion tensor (DTI) image showing the trigeminal nerve root entry zone in a single participant. The DTI image is color-coded for direction of greatest water movement. The outline of the trigeminal nerve region used for the total nerve analysis is shown in yellow shading on the T1-weighted anatomical image and is also outlined in yellow on the DTI images. This volume of interest encompassing the root entry zone is the section of the trigeminal nerve that lies within the pontine cistern, i.e., from where it emerges from the pons, to the point at which it exits the pontine cistern anteriorly. Plots of FA left nerve values for individual controls and migraineurs are shown in the right inset. Horizontal bars indicate mean (±SEM) in the control and migraine groups.

Figure 2

Areas of the brainstem in which reductions in trigeminal nerve fractional anisotropy (FA) were significantly correlated with increases in brainstem mean diffusivity (MD) assessed in migraine (n = 32) and control participants (n = 31). Significant clusters are overlaid onto a mean T1-weighted brainstem template image. Slice locations in Montreal Neurological Institute (MNI) space are indicated to the top right of each slice. Note that MD was significantly correlated to nerve FA in the brainstem regions of the spinal trigeminal nucleus (SpV), dorsolateral pons (dlPons) and the periaqueductal gray matter (PAG).

Figure 3

(A) Plots of mean diffusivity (MD), axial diffusivity (AX) and radial diffusivity (RD) against trigeminal nerve fractional anisotropy (FA) for the spinal trigeminal nucleus (SpV), dorsolateral pons (dlPons) and the periaqueductal gray matter (PAG). In addition, to the right of each correlation plot are plots of mean (±SEM) MD, AX and RD values for each of the significant brainstem clusters. (B) Plots of gray matter volume against trigeminal nerve FA derived from significant MD clusters and plots of mean (±SEM) gray matter volume for each of the significant MD brainstem clusters.

Figure 4

Areas above the brainstem in which reductions in trigeminal nerve fractional anisotropy (FA) were significantly correlated with increases in above the brainstem mean diffusivity (MD) assessed in migraine (n = 38) and control participants (n = 38). Significant clusters are overlaid onto a mean T1-weighted anatomical image. Slice locations in Montreal Neurological Institute (MNI) space are indicated to the top right of each slice. Note that MD was significantly correlated to nerve FA in the posterior cingulate cortex (PCC), primary somatosensory cortex (S1), dorsoposterior insula (dp insula), hypothalamus, putamen and the primary visual cortex (V1).

Figure 5

(A) Plots of mean diffusivity (MD), axial diffusivity (AX) and radial diffusivity (RD) against trigeminal nerve fractional anisotropy (FA) for the posterior cingulate cortex (PCC), primary somatosensory cortex (S1), dorsoposterior insula (dp insula), hypothalamus, putamen and primary visual cortex (V1). In addition, to the right of each correlation plot are plots of mean (±SEM) MD, AX and RD values for each of the significant above the brainstem clusters. (B) Plots of gray matter volume against trigeminal nerve FA derived from significant MD clusters and plots of mean (±SEM) gray matter volume for each of the significant MD above the brainstem clusters.