Dynamic functional connectivity changes in noise-induced hearing loss: a resting-state fMRI study with machine learning-based classification

Motivation: Noise-induced hearing loss (NIHL) is the second most common type of hearing loss among adults. Goal(s): To utilise dynamic functional connectivity (dFC) analysis in NIHL. Approach: 58 NIHL and 42 healthy controls (HCs) underwent resting-state functional magnetic resonance imaging (GE discovery 750). The sliding window approach was employed to evaluate dFC between region pairs. These features were used to construct support vector machine (SVM) classifiers. Results: Compared with HCs, NIHL demonstrated decreased dFC between the right supplementary motor (SMA.R) and bilateral cuneus, increased dFC between the SMA.R and left inferior parietal. The accuracy of SVM classifier based on abnormal dFC features (FDR<0.05) 82.5%. Impact: We applied a sliding window dFC method to examine the whole-brain activity in NIHL, and employed machine learning techniques to construct a support vector machine model. Our findings provide a better understanding of the brain network alterations associated with NIHL.