Abstract
BACKGROUND: The thalamo-prefrontal white matter (WM) pathway, a core structural element of the frontal-limbic system disrupted in premenstrual syndrome (PMS), remains poorly understood. METHODS: Diffusion tensor imaging (DTI), functional MRI (fMRI), and serum cytokine levels were collected from 41 PMS participants and 51 healthy controls (HCs), all diagnosed using the Daily Record of Severity of Problems (DRSP) scale. Bilateral thalamic-frontal WM pathways-the anterior thalamic radiations (ATRs)-were reconstructed using probabilistic fiber tracking. Two-sample tests examined group differences in fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), and amplitude of low-frequency fluctuation (ALFF) within bilateral ATRs. Spearman correlations assessed associations among these MRI metrics, inflammatory cytokines, and DRSP scores. Machine learning models further evaluated the diagnostic and predictive utility of left ATR features combined with inflammatory cytokines. RESULTS: Compared to HCs, PMS patients exhibited increased MD, AD, RD, and ALFF values in the left ATR, as well as elevated tumor necrosis factor (TNF)-α levels. Correlation analysis revealed that these MRI alterations in the left ATR and TNF-α levels were linked to DRSP scores. Additionally, the machine learning models constructed using the optimal feature subset, involved in MD, AD and ALFF of left ATR as well as TNF-α, demonstrated robust performance in diagnosing PMS and predicting DRSP scores. CONCLUSION: These findings suggest altered thalamo-frontal WM connectivity and elevated TNF-α in PMS. The left ATR may serve as a biomarker of PMS neuro-mechanisms when combined with multi-MRI and inflammation metrics.