Altered Coupling of Psychological Relaxation and Regional Volume of Brain Reward Areas in Multiple Sclerosis

Background: Psychological stress can influence the severity of multiple sclerosis (MS), but little is known about neurobiological factors potentially counteracting these effects. Objective: To identify gray matter (GM) brain regions related to relaxation after stress exposure in persons with MS (PwMS). Methods: 36 PwMS and 21 healthy controls (HCs) reported their feeling of relaxation during a mild stress task. These markers were related to regional GM volumes, heart rate, and depressive symptoms. Results: Relaxation was differentially linked to heart rate in both groups (t = 2.20, p = 0.017), i.e., both markers were only related in HCs. Relaxation was positively linked to depressive symptoms across all participants (t = 1.99, p = 0.045) although this link differed weakly between groups (t = 1.62, p = 0.108). Primarily, the volume in medial temporal gyrus was negatively linked to relaxation in PwMS (t = −5.55, pfamily−wise−error(FWE)corrected = 0.018). A group-specific coupling of relaxation and GM volume was found in ventromedial prefrontal cortex (VMPFC) (t = −4.89, pFWE = 0.039). Conclusion: PwMS appear unable to integrate peripheral stress signals into their perception of relaxation. Together with the group-specific coupling of relaxation and VMPFC volume, a key area of the brain reward system for valuation of affectively relevant stimuli, this finding suggests a clinically relevant misinterpretation of stress-related affective stimuli in MS.


Statistical analyses of MRI group data
Supplementary VBM analysis 1: Analysis of differences in regional GM volume between

PwMS and HCs
In order to identify areas with smaller regional GM volume in PwMS than HCs, we conducted a regression analysis using SnPM13. Group membership (0: HC, 1, PwMS) served as CI, sex, age, the participants intracranial volume and a constant served as CNI. Exactly as all other VBM MRI analyses, this analysis was constrained to coordinates located in the GM group mask. In this explorative analysis, we report the MNI-coordinates for the voxel with most extreme t-statistic (i.e. for the peak voxel) in a cluster of voxels reaching significance following a multiple comparison or family-wise error (FWE) corrected significance threshold (αFWE = 0.1) for negative effects.
Supplementary VBM analysis 2: Anatomically constrained analyses of associations of perceived relaxation and regional GM volume in PwMS and of differential association of perceived relaxation and regional GM volume in PwMS and HCs In the second supplementary VBM analysis, we repeated the two VBM analyses described in the main text. The only difference between the analyses presented in the main text and this supplementary one was, that we exclusively searched across voxel coordinates identified as significantly affected by GM loss in PwMS in the above supplementary analysis in the present supplementary analyses.
Supplementary VBM analysis 3: Analysis of associations of perceived relaxation and regional

GM volume in HCs
In this supplementary analysis, we tested associations between regional GM volume and immediate and delayed relaxation in the 21 HCs. Except for the uncorrected false positive threshold used for inference (α = 0.001), all aspects were as described in the main text for the corresponding analysis in PwMS. Given that the primary aim of this analysis was to enable a comparison of main effects of relaxation on regional GM volume in HCs to those observed in PwMS, the more liberal uncorrected significance threshold of α = 0.001 was chosen to compensate for the loss of statistical power in the HC analysis following from the smaller number of HCs compared to PwMS.
In a follow-up step, we evaluated whether the coordinates found in the first VBM analysis presented in the main text in PwMS (exclusively showing negative associations between GM volume and relaxation) did spatially overlap with any of the coordinates found in supplementary VBM analysis 3 (irrespective of the directionality of associations found in the supplementary analysis and the type of relaxation [i.e., immediate and delayed]).
Supplementary VBM analysis 4: Analysis of associations between relaxation and regional GM volume in PwMS additionally considering the putative impact of progressive MS To test whether the presence of (secondary) progressive MS (vs. RRMS) has an impact on the link between perceived relaxation and regional GM volume in MS, we repeated the corresponding analyses presented in the main text but now included a regressor coding ones for SPMS and zeros for RRMS as additional CNI. This was done separately for immediate and for delayed relaxation. Due to the inclusion of an additional CNI, we report the MNI-coordinates for the voxel with most extreme t-statistic (i.e. for the peak voxel) in a cluster of voxels reaching significance following a multiple comparison or family-wise error (FWE) corrected significance threshold of αFWE = 0.1.

Statistical analyses of MRI group data
Supplementary VBM analysis 1: Analysis of differences in regional GM volume between

PwMS and HCs
The below supplementary Table S1 reports coordinates with significant less GM volume in PwMS than HCs. Supplementary VBM analysis 2: Anatomically constrained analyses of associations of perceived relaxation and regional GM volume in PwMS and of differential association of perceived relaxation and regional GM volume in PwMS and HCs These supplementary analyses testing associations of perceived relaxation and regional GM volume in PwMS and differential association of perceived relaxation and regional GM volume in PwMS and HCs in brain areas significantly affected by GM loss in PwMS did not reveal any significant results.

Supplementary VBM analysis 3: Analysis of associations of perceived relaxation and regional GM volume in HCs
The below supplementary tables report GM coordinates positively or negatively related to regional GM volume in HCs in terms of immediate relaxation (Table S2) or delayed relaxation respectively (Table S3). x, y, z: coordinates in MNI space; t: t-statistic; puncorr.: uncorrected probability for observing the given t-value by chance on the voxel level (α = 0.001). pFWE: family-wise-error corrected probability for observing the given t-value by chance on the voxel level. Please note, that the non-monotonic increase in uncorrected p-values across t-statistics with decreasing absolute magnitude results from the permutation method applied by SnPM13. Supplementary VBM analysis 4: Analysis of associations between relaxation and regional GM volume in PwMS additionally considering the putative impact of progressive MS The below supplementary Table S4 reports coordinates with an association between relaxation measures and regional GM volume in PwMS according to a regression model including MS type (0 -RRMS, 1 -SPMS) as additional CNI.