Classical trigeminal neuralgia (CTN) is a common and severe chronic neuropathic facial pain disorder. The pathological mechanisms of CTN are not fully understood. Recent studies have shown that resting-state functional magnetic resonance imaging (rs-fMRI) could provide insights into the functional changes of CTN patients and the complexity of neural processes. However, the precise spatial pattern of complexity changes in CTN patients is still unclear. This study is designed to explore the spatial distribution of complexity alterations in CTN patients using brain entropy (BEN).
A total of 85 CTN patients and 79 age- and sex-matched healthy controls (HCs) were enrolled in this study. All participants underwent rs-fMRI and neuropsychological evaluations. BEN changes were analyzed to observe the spatial distribution of CTN patient complexity, as well as the relationship between these changes and clinical variables. Sixteen different machine learning methods were employed to classify the CTN patients from the HCs, and the best-performing method was selected.
Compared with HCs, CTN patients exhibited increased BEN in the thalamus and brainstem, and decreased BEN in the inferior semilunar lobule. Further analyses revealed a low positive correlation between the average BEN values of the thalamus and neuropsychological assessments. Among the 16 machine learning methods, the Conditional Mutual Information Maximization-Random Forest (CMIM-RF) method yielded the highest area under the curve (AUC) of 0.801.
Our study demonstrated that BEN changes in the thalamus and pons and inferior semilunar lobule were associated with CTN and machine learning methods could effectively classify CTN patients and HCs based on BEN changes. Our findings may provide new insights into the neuropathological mechanisms of CTN and have implications for the diagnosis and treatment of CTN.