Multilayer network analysis of dynamic network reconfiguration in age-related hearing loss: a cross-sectional
Jiajie Song1, Shuo Li2, Jun Yao2
1Department of Radiology, Nanjing Pukou People's Hospital, Liangjiang Hospital, Southeast University, Nanjing, China.
Background:
The neural mechanisms associated with age-related hearing loss (ARHL) and their broader implications for brain function remain incompletely understood. This study explored the dynamic network reconfiguration in ARHL using dynamic graph theory and multilayer network analysis.
Methods:
Resting-state functional magnetic resonance imaging (MRI) assessments were conducted on 62 patients with ARHL and 58 healthy controls (HCs) matched for age (≥60 years), sex (ARHL: 32 males, 30 females; HCs: 30 males and 28 females), and education level (all at least 8 years). Cognitive performance was comprehensively assessed in both groups with a battery of standardized neuropsychological tests. Dynamic brain functional networks were analyzed via graph theory to investigate local and global network metrics. Multilayer network analysis was employed to identify changes in global brain network exchanges in ARHL. Spearman correlation analysis was used to calculate the relationship between functional MRI data and cognitive scores.
Results:
No demographic differences were observed between patients with ARHL and HCs (P>0.05). However, patients with ARHL exhibited significantly poorer pure-tone audiometry results (P<0.001) and inferior performance on both Trail Making Test-B and Complex Figure Test delayed tests (P<0.05). The ARHL group exhibited reduced local efficiency (P=0.01) and increased node efficiency (P=0.00054). Significant differences in network-switching rates were observed in the left orbital middle frontal gyrus (P=0.006), right frontal orbital inferior (P=0.032), right olfactory cortex (P=0.010), left medial superior frontal gyrus (P=0.045), right middle occipital gyrus (P=0.021), right superior parietal gyrus (P=0.008), left inferior parietal lobule (P=0.034), right caudate nucleus (P=0.020), left globus pallidus (P=0.007), and left superior temporal pole (TPOsup.L) (P=0.022). The network-switching rate of the left TPOsup in the ARHL group was positively correlated with cognitive scores (r=0.270; P=0.034).
Conclusions:
Based on dynamic graph theory and multilayer network analysis, abnormal dynamic network reconfiguration in patients with ARHL were revealed, with reduced network-switching rates in several brain regions. These findings highlight the functional significance of network-switching rates and provide new insights into the neural mechanisms associated with ARHL.
