Abnormal White Matter Functional-Structural Coupling in Noise-Induced Hearing Loss: A Multimodal Study Combining...

To investigate abnormal coupling patterns between white matter functional activation (static and dynamic amplitude of low‑frequency fluctuations, sALFF and dALFF) and white matter (WM) microstructural integrity (fractional anisotropy, FA) in patients with noise‑induced hearing loss (NIHL).

Fifty‑three patients with NIHL and 50 matched healthy controls (HCs) underwent resting‑state fMRI and diffusion tensor imaging. sALFF and dALFF were computed to characterize local neural activity in WM; WM microstructural integrity was assessed using tract‑based spatial statistics (TBSS) to generate a FA skeleton; WM functional-structural coupling was quantified by voxel‑wise ratios (ALFF/FA) on the skeleton; the overlapping regions relationship between ALFF and FA was further examined by correlation analysis (p < 0.05, Gaussian random field (GRF) correction). Significant abnormal imaging features and blood biomarkers were selected via least absolute shrinkage and selection operator (LASSO) to construct a support vector machine (SVM) classifier.

NIHL patients exhibited widespread FA reductions across multiple white matter tracts, with a single FA increase in the right anterior corona radiata. On the FA skeleton, sALFF was increased in the left middle cerebellar peduncle (MCP), right inferior longitudinal fasciculus (ILF), and left cingulum hippocampus, whereas dALFF was decreased in the left ILF and corpus callosum trunk. Voxel‑wise ratios revealed: (i) increased sALFF/individual FA in the right uncinate fasciculus and right cingulum bundle, and decreased sALFF/individual FA in the left fronto‑insular tract and left cingulum bundle; (ii) increased dALFF/individual FA in the right fronto‑insular tract, and decreased dALFF/individual FA in the left corticostriate fibers (all p < 0.05, GRF‑corrected). In overlapping regions where both FA and ALFF showed significant group differences, correlation analysis showed that sALFF tended to be negatively correlated with FA in NIHL patients, whereas dALFF showed a positive correlation trend - opposite to the pattern observed in HCs. Clinically, the sALFF/FA ratio in the overlapping left MCP region negatively correlated with noise exposure time (r = -0.330, p < 0.001), whereas the sALFF/FA ratio in the left cingulum hippocampus positively correlated with anxiety scores (r = 0.238, p = 0.015), and the dALFF/FA ratio in the left ILF1 positively correlated with hearing threshold (r = 0.311, p = 0.023). A LASSO‑based SVM model using six multimodal features discriminated NIHL from controls with 84.5% accuracy and an AUC of 0.893 (95% CI 0.83 - 0.95).

This study demonstrates widespread white matter abnormalities and functional-structural decoupling in NIHL, characterized by region‑specific alterations in sALFF/FA and dALFF/FA. These decoupling patterns correlate with key clinical variables (noise exposure duration, anxiety, and hearing threshold), supporting the dysconnectivity hypothesis and offering potential imaging biomarkers.