Altered effective connectivity in leucine-rich glioma-inactivated 1 antibody encephalitis: a spectral dynamic causal modeling study.

Background: Despite advances in understanding the effective connectivity (EC) of brain networks in leucine-rich glioma-inactivated 1 (LGI1) antibody encephalitis, the specific cause and underlying mechanisms of LGI1 encephalitis remain unclear. Materials and methods: The study included 27 patients with anti-LGI1 encephalitis and 28 age- and sex-matched normal controls. Amplitude of low-frequency fluctuation (ALFF) analysis identified altered brain regions. Spectral dynamic causal modeling (spDCM) then assessed EC between these regions. Relationships between EC strength and both clinical severity and cognitive function were analyzed. Results: Distinct EC patterns were found in patients versus controls. Specifically, inhibitory EC was observed from the hippocampus to the superior temporal gyrus, while excitatory EC was noted in the reverse direction. Patients also showed reduced inhibitory self-connections in the posterior cingulate cortex. Crucially, inhibitory EC from the right hippocampus to the left superior temporal gyrus correlated inversely with symptom severity and positively with cognitive performance. Conversely, reduced inhibitory self-connections in the posterior cingulate cortex correlated positively with symptom severity and negatively with cognitive function. Conclusions: These findings indicate that changes in causal connections between specific brain regions significantly contribute to neurological deficits in anti-LGI1 encephalitis. The inhibitory connectivity from the hippocampus to the superior temporal gyrus may serve as a potential biomarker for personalized diagnosis, offering new insights into the underlying pathological mechanisms of this disorder.