Neurovascular Coupling During Visual Stimulation in Multiple Sclerosis: A MEG-fMRI Study

Abstract

The process of neurovascular coupling ensures that increases in neuronal activity are fed by increases in cerebral blood flow. Evidence suggests that neurovascular coupling may be impaired in Multiple Sclerosis (MS) due to a combination of brain hypoperfusion, altered cerebrovascular reactivity and oxygen metabolism, and altered levels of vasoactive compounds. Here, we tested the hypothesis that neurovascular coupling is impaired in MS. We characterized neurovascular coupling as the relationship between changes in neuronal oscillatory power within the gamma frequency band (30-80 Hz), as measured by magnetoencephalography (MEG), and associated hemodynamic changes (blood oxygenation level dependent, BOLD, and cerebral blood flow, CBF) as measured by functional MRI. We characterized these responses in the visual cortex in 13 MS patients and in 10 matched healthy controls using a reversing checkerboard stimulus at five visual contrasts. There were no significant group differences in visual acuity, P100 latencies, occipital gray matter (GM) volumes and baseline CBF. However, in the MS patients we found a significant reduction in peak gamma power, BOLD and CBF responses. There were no significant differences in neurovascular coupling between groups, in the visual cortex. Our results suggest that neuronal and vascular responses are altered in MS. Gamma power reduction could be an indicator of GM dysfunction, possibly mediated by GABAergic changes. Altered hemodynamic responses confirm previous reports of a vascular dysfunction in MS. Despite altered neuronal and vascular responses, neurovascular coupling appears to be preserved in MS, at least within the range of damage and disability studied here.

Keywords: Multiple Sclerosis; cerebral blood flow; functional MRI; magnetoencephalography; neurovascular coupling; visual function.

Fig. 1

Method used to characterize the peak gamma power response, for each participant. After performing the Hilbert transform on each trail, and averaging over trials, we obtain power changes in the time–frequency domain, relative to baseline. Shown on the left is a time–frequency plot for one participant, at one visual contrast (50%). For each 250-ms epochs (corresponding to one checkerboard reversal), we average over time. We then extract the peak amplitude change from baseline (on the right) which is shown in this example to be approximately 15%, at 40 Hz (indicated by the red arrow). We take an average of the four values (one from each of the four 250-ms epochs) to give the final peak gamma power change from baseline.

Fig. 2

The location of the ROI used to extract the BOLD and CBF responses (yellow) for every participant, overlaid on the primary visual cortex (red). The top plot shows the left eye analysis and the bottom plot shows the right eye analysis. The dots indicate the location of the peak voxel (peak of the gamma response, percentage change from baseline) for each individual participant (blue = controls, green = MS patients), which was used for the time–frequency analysis of the MEG data. This is shown for n = 10 controls n = 13 MS patients.

Fig. 3

Beamformer contrast images (band-pass filtered 30–80 Hz) measured as percentage change between stimulus and baseline, projected onto a template brain surface. This is shown for controls (left column) and MS patients (middle column). The right column illustrates the t-statistic values for the difference between patients and controls (negative values indicating lower amplitude for patients than controls). Simply for illustration purposes, t-values are plotted here at the uncorrected level. The data were averaged over both eyes. Only right medial and lateral views are shown; the same trends were seen for left.

Fig. 4

Significant CBF voxels at the group level in response to the visual checkerboard stimulus, compared between MS patients and controls. These data are an average of both eyes. The significant activity shown is the average activity across all visual stimulus conditions. Voxels were thresholded using clusters determined by Z > 2.3 and a (corrected) cluster significance threshold of p = 0.05 (Worsley, 2001). The bottom plot shows voxels that showed significantly greater activity in the patient group, compared with controls. This activity was localized to the intracalcarine and supracalcarine cortex, as well as the cuneus (using Harvard-Oxford Cortical Structural Atlas). There were no voxels showing significantly greater activity for the control group compared to the MS patients. The right side of the image corresponds to the left side of the brain.

Fig. 5

Significant BOLD voxels at the group level in the visual checkerboard stimulus, compared between MS patients and controls. These data are an average of both eyes. The significant activity shown is the average activity across all visual stimulus conditions. Voxels were thresholded using clusters determined by Z > 2.3 and a (corrected) cluster significance threshold of p = 0.05 (Worsley, 2001). The bottom plot shows voxels that showed significantly greater activity in the patient group, compared with controls. This activity was localized to the lingual gyrus, intracalcarine cortex, pre-cuneus and cuneus (using Harvard-Oxford Cortical Structural Atlas). There were no voxels showing significantly greater activity for the control group compared to the MS patients. The right side of the image corresponds to the left side of the brain.

Fig. 6

The effect of group and visual contrast on peak gamma power, BOLD signal, CBF signal (percentage change from baseline) and CBF signal (ml/100 g/min change from baseline). The graph displays Mean ± SEM. For BOLD, CBF and CBF quantified the effect of group was not significantly different across contrast level, so the p values refer to the main effect. For Gamma, there was an interaction between group and contrast level so the p values refer to the simple main effects. ^*p < 0.05, ^**p < 0.001. All pairwise comparisons between contrast levels were significant at an alpha level of 0.05, with Holm–Bonferroni correction.

Fig. 7

(A) shows the relationship between peak gamma power change and BOLD, CBF and CBF changes (quantified) in response to the visual checkerboard stimulus. Group median values are plotted. Each point represents a different contrast level for each eye. The reduced range of electrophysiological and the hemodynamic responses are evident in the patient group. (B) The relationship between peak gamma power change and BOLD change shown for each control (n = 9) and each patient (n = 12) separately. The different colors represent the different participants and the black lines the linear model fit.