PURPOSE

The ability to distinguish myelin oligodendrocyte glycoprotein antibody-seropositive optic neuritis (MOG-ON) from seronegative-ON is critical in clinical practice. We investigate potential neural mechanisms and differentiation biomarkers via large-scale functional network connectivity (FNC) using resting-state functional magnetic resonance imaging (RS-fMRI).

METHODS

RS-fMRI-based independent component analysis (ICA) was performed in 79 subjects, including 23 with MOG-ON, 30 with seronegative-ON and 26 healthy controls (HCs). The resting-state networks (RSNs) extracted from the ICA were used to investigate static FNC (sFNC) changes within and between groups. In addition, 5 dynamic FNC (dFNC) states were identified using k-means cluster analysis, and several state-related properties were calculated. Receiver operating characteristic (ROC) curve analysis was also performed to determine its value in differential diagnosis.

RESULTS

In the sFNC analysis, the patient groups showed decreased intranetwork functional connectivity (FC) within several RSNs compared to the HC group. The MOG-ON group presented significantly altered intranetwork FC in the medial visual network (mVN) and posterior default mode network (pDMN) compared with the seronegative-ON group. Compared with the HCs, the patient groups also presented abnormal internetwork FC between RSNs. In the dFNC analysis, the patient groups presented altered fractional occupancy and dwell times in states 1 and 5 compared with HCs, and the changes in state-related metrics were also distinct between the MOG-ON and seronegative-ON groups. In terms of ROC curve analysis, optimal diagnostic performance was achieved by combining static and dynamic approaches.

CONCLUSIONS

Abnormal large-scale static and dynamic brain functional networks may help to better understand the neural mechanisms of MOG-ON and seronegative-ON and their differentiation.