Introduction

Multiple sclerosis (MS) is a serious, progressive, and disabling disease, characterized by inflammatory demyelination in the central nervous system. The gender effects have been observed in patients with MS and affect the prevalence¹, course², disability³ and responses to therapy⁴. Recently, large cohort neuroimaging studies showed that MS males seem to have a higher level of brain damage compared to females^5,6. These findings supported the idea that gender-related effects may affect brain structure in MS patients. However, it is not yet clear whether gender-related factors affect functional network connectivity in MS. Thus, our study aims to investigate gender-related differences in static and dynamic connectivity within large scale functional networks in MS, and the associations with structural and clinical measures.

Materials and methods

A total of 208 patients with MS (135 females / 73 males) and 228 healthy controls (HC; 123 females / 105 males) were recruited from 7 Chinese centers for this retrospective study. All subjects underwent 3.0T scanner, and obtained high resolution 3D T1 weight, T2 weight, fluid-attenuated inversion recovery, and functional MR images. And physical disability of MS patients was measured by using the expanded disability status scale (EDSS). To define resting state networks, spatial independent component analysis was implemented using the Group ICA Of fMRI Toolbox, then obtained 20 components and 7 networks (Figure 1). To access the intra- and inter-network connectivity, spatial map intensity and functional network connectivity were analyzed. Moreover, the sliding window and k-means clustering methods were used to evaluate the connectivity pattern and variability of dynamic functional network connectivity (dFNC). Additionally, structural measurements were calculated, including T2 lesion load, total intracranial volume, white matter volume (WMV) and gray matter volume (GMV). Then, one-way ANOVA tests were used to assess the difference of static and dynamic connectivity at the group level (Healthy females vs. Healthy males, MS females vs. MS males, MS females vs. Healthy females, MS males vs. Healthy males; FDR corrected, P <0.05). Finally, partial correlations were used to analyze the associations between functional networks connectivity and structural, or clinical characteristics (P < 0.05).

Results

Demographic data and clinical characteristics are shown in Table 1 (Figure 2). In patients with MS, male patients seem to show younger ages (35.42±10.98 years), longer disease duration (median: 24 months), and higher disability (median EDSS: 2.5). Statistical analyses revealed that there is no significant difference in age, disease duration and EDSS score between female and male patients (P > 0.10). Moreover, MS males have more lesion load (P = 0.041, uncorrected) compared to MS females.
In functional network analyses, widespread differences within intra- and inter-network static functional connectivity were observed between healthy males and females: stronger connectivity in females than males (Figure 3A, 3B). But these gender-related differences were significantly eliminated in MS patients. When compared to HC, the differences of static connectivity were observed in the male patients, not in female patients (Figure 3C, 3D), particularly decreased connectivity within ventral attention network, and between sensorimotor network and visual network (FDR correlated, P < 0.05).
After cluster analysis, three dFNC states were identified for all participants: State 1 (weakly connected state; 49%), State 2 (middle-high connected state; 36%) and State 3 (high connected state; 15%) (Figure 4A). Healthy females preferred spending more time in middle-high connected state with fewer transitions, while healthy males and MS patients tended to stay in weakly connected state with frequent transitions (Figure 4C). Group-level analyses suggested gender-related dFNC differences in HC, which disappeared in MS. Most importantly, compared with healthy females, MS females showed more decreased connectivity in all states (Figure 5). While no dynamic significance between healthy and MS males. The dFNC results were corrected by FDR (P < 0.05).
Correlation analyses suggested functional connectivity within frontoparietal network (r = -0.251, P = 0.044) and between frontoparietal network and sensorimotor network (r = -0.311, P = 0.012) were negatively correlated with GMV in MS males. In MS females, EDSS scores (r = 0.216, P = 0.015) and GMV (r = -0.186, P = 0.037) were correlated with dell time in State 1, GMV (r = -0.291, P = 0.001) and WMV (r = -0.252, P = 0.004) were observed to be negatively related to the numbers of transition.

Discussion

Our results revealed that MS could eliminate the gender-related effect in healthy controls, and only male patients showed decreased functional connectivity and higher lesion load compared to HC. Our findings confirmed and further supported the idea that male patients have a higher degree of brain damage from functional evidence^6,7. Contrary to static observations, decreased dFNC was displayed only in female patients. In addition, MS females tended to stay in weakly connected state, and the dell time is related to disability and GMV, which indicated that the dFNC damage occurred accompanying structural damage and clinical disability in MS females. Finally, MS females have higher number of transitions, suggesting that MS may impair the stability of dFNC patterns in females.

Conclusion

Our results demonstrated that MS mainly affects static connectivity in males and dynamic connectivity in females, suggesting that gender-related effects may play an important role in functional impairment and reorganization of MS, and be related to clinical disability.

Acknowledgements

We are grateful for support and discussions from our colleagues in other centers.