Functional connectivity in the dorsal network of the cervical spinal cord is correlated with diffusion tensor imaging indices in relapsing-remitting multiple sclerosis

Abstract

Focal lesions may affect functional connectivity (FC) of the ventral and dorsal networks in the cervical spinal cord of people with relapsing-remitting multiple sclerosis (RRMS). Resting-state FC can be measured using functional MRI (fMRI) at 3T. This study sought to determine whether alterations in FC may be related to the degree of damage in the normal-appearing tissue. Tissue integrity and FC in the cervical spinal cord were assessed with diffusion tensor imaging (DTI) and resting-state fMRI, respectively, in a group of 26 RRMS participants with high cervical lesion load, low disability, and minimally impaired sensorimotor function, and healthy controls. Lower fractional anisotropy (FA) and higher radial diffusivity (RD) were observed in the normal-appearing white matter in the RRMS group relative to controls. Average FC in ventral and dorsal networks was similar between groups. Significant associations were found between higher FC in the dorsal sensory network and several DTI markers of pathology in the normal-appearing tissue. In the normal-appearing grey matter, dorsal FC was positively correlated with axial diffusivity (AD) (r = 0.46, p = 0.020) and mean diffusivity (MD) (r = 0.43, p = 0.032). In the normal-appearing white matter, dorsal FC was negatively correlated with FA (r = -0.43, p = 0.028) and positively correlated with RD (r = 0.49, p = 0.012), AD (r = 0.42, p = 0.037) and MD (r = 0.53, p = 0.006). These results suggest that increased connectivity, while remaining within the normal range, may represent a compensatory mechanism in response to structural damage in support of preserved sensory function in RRMS.

Keywords: Diffusion tensor imaging; Functional connectivity; Multiple sclerosis; Resting-state fMRI; Spinal cord.

Fig. 1

Example data for one control (25-year-old female) and one participant with MS (25-year-old female, EDSS 0, disease duration < 1 year). Shown are the mFFE anatomical image used for lesion identification, vertebral level identification and segmentation; the average functional image showing grey/white matter contrast after motion correction; grey matter horns regions of interest in functional space used to derive functional correlations; tissue type segmentations in diffusion space overlaid on a b = 0 image; and FA, RD, AD and MD quantitative maps. A dorsal column lesion in the MS participant is shown with a red arrow, and the corresponding lesion mask in DTI space in red on the segmentation image. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

(A) Functional connectivity indices between groups. Ventral and dorsal network z-scores do not significantly differ between control and MS participants. (B) Distribution of z-scores by EDSS in MS participants compared to controls’ values. Note the EDSS scale is not continuous. One dorsal z-score is missing for one MS participant (23-year-old female) due to data quality.

Fig. 3

Comparison of DTI values in GM and WM between controls, MS normal-appearing and lesioned tissue. †Trend level result (p < 0.10, non-significant). Significant at *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 4

Correlations between normal-appearing grey and white matter DTI indices and dorsal (sensory) z-scores in the MS group. Pearson’s r correlation coefficients are reported. Line and shaded area represent linear regression line and 95% confidence intervals. Significant at *p < 0.05, **p < 0.01.