The relationship between changes in functional connectivity gradients and cognitive-emotional disorders in sudden sensorineural hearing loss

Abstract

Sudden sensorineural hearing loss, a prevalent emergency in otolaryngology, is known to potentially precipitate cognitive and emotional disorders in affected individuals. Extensive research has documented the phenomenon of cortical functional reorganization in patients with sudden sensorineural hearing loss. However, the potential link between this neural functional remodelling and cognitive-emotional disorders remains unclear. To investigate this issue, 30 bilateral sudden sensorineural hearing loss patients and 30 healthy adults were recruited for this study. We collected clinical data and resting-state functional magnetic resonance imaging data from the participants. Gradient mapping analysis was employed to calculate the first three gradients for each subject. Subsequently, gradient changes in sudden sensorineural hearing loss patients were compared with healthy controls at global, regional and network levels. Finally, we explored the relationship between gradient values and clinical variables. The results revealed that at the global level, sudden sensorineural hearing loss did not exhibit significant differences in the primary gradient but showed a state of compression in the second and third gradients. At the regional level, sudden sensorineural hearing loss patients exhibited a significant reduction in the primary gradient values in the temporal pole and ventral prefrontal cortex, which were closely related to neuro-scale scores. Regarding the network level, sudden sensorineural hearing loss did not show significant differences in the primary gradient but instead displayed significant changes in the control network and default mode network in the second and third gradients. This study revealed disruptions in the functional hierarchy of sudden sensorineural hearing loss, and the alterations in functional connectivity gradients were closely associated with cognitive and emotional disturbances in patients. These findings provide new evidence for understanding the functional remodelling that occurs in sudden sensorineural hearing loss.

Keywords: cognitive–emotional disorders; functional connectivity gradients; functional magnetic resonance imaging; sudden sensorineural hearing loss.

Graphical Abstract

Figure 1

Connectome gradient mapping for the SSNHL and HC groups. (A) The distribution of individual-averaged principal gradient scores on the cortical surface in individuals with SSNHL and HC. The global histogram of principal gradient related to the 400 regions in individuals with SSNHL and HC. (B) The distribution of individual-averaged secondary gradient scores on the cortical surface in individuals with SSNHL and HC. The global histogram of the secondary gradient related to the 400 regions in individuals with SSNHL and HC. (C) The distribution of individual-averaged third gradient scores on the cortical surface in individuals with SSNHL and HC. The global histogram of the third gradient related to the 400 regions in individuals with SSNHL and HC. SSNHL, sudden sensorineural hearing loss; HC, healthy control.

Figure 2

The scatter plot of the first three gradient scores in the SSNHL and HC groups. (A) The scatter plot of the first three gradient scores in patients with SSNHL. (B) The scatter plot of the first three gradient scores in patients with HC. SSNHL, sudden sensorineural hearing loss; HC, healthy control.

Figure 3

The functional network gradient comparisons in principal gradient between SSNHL (n = 30) and HC groups (n = 30) assessed by two-sample t-test. (A) First gradient values distribution in seven functional networks. (B) The region-level differences of the principal gradient between two groups (two-sample t-test, P < 0.05, FDR uncorrected). The heat map scale bar displays the range of t-values for distinct brain regions, where the magnitude of these values reflects the degree of between-group differences in each region. Larger absolute values indicate greater between-group differences in that brain region. (C) The network-level differences of the principal gradient between two groups. Through the two-sample t-test, no significant group differences were observed at the network level (two-sample t-test, P = 0.05, FDR uncorrected). SSNHL, sudden sensorineural hearing loss; HCs, healthy controls; DMN, default mode network; CON, control; LIM, limbic; SAL, salience attention; DAN, dorsal attention; SOM, somatomotor; VIS, visual; TP, temporal pole; vPFC, ventral prefrontal cortex.

Figure 4

The local gradient differences in principal gradient between SSNHL (n = 30) and HC groups (n = 30) assessed by two-sample t-test. (A). Group differences in the TP temporal pole that had significance differences (two-sample t-test, FDR-corrected P < 0.05). In the TP temporal pole, the gradient values in SSNHL were significantly lower than those in HC. (B) Group differences in the vPFC that had significance differences (two-sample t-test, FDR-corrected P < 0.05). In the vPFC, the gradient values in SSNHL were significantly lower than those in HC. (C) The differences of the gradient range between patients with SSNHL and HC (two-sample t-test, P > 0.05). (D) The differences of the gradient variance between patients with SSNHL and HC (two-sample t-test, P > 0.05). (E) The differences of the gradient explained ratio between patients with SSNHL and HC (two-sample t-test, P > 0.05). SSNHL, sudden sensorineural hearing loss; HCs, healthy controls; DMN, default mode network; CON, control; LIM, limbic; SAL, salience attention; DAN, dorsal attention; SOM, somatomotor; VIS, visual; TP, temporal pole; vPFC, ventral prefrontal cortex. ***P < 0.005; ns, no significant.

Figure 5

The gradient comparisons in the second gradient between SSNHL (n = 30) and HC groups assessed by independent samples t-tests. (A) The region-level differences of the principal gradient between two groups (two-sample t-test, P < 0.05, FDR uncorrected). The heat map scale bar displays the range of t-values for distinct brain regions, where the magnitude of these values reflects the degree of between-group differences in each region. Larger absolute values indicate greater between-group differences in that brain region. (B) Group differences in the second gradient values at the network level (two-sample t-test, *FDR-uncorrected P < 0.05). SSNHL, sudden sensorineural hearing loss; HC, healthy control.

Figure 6

The gradient comparisons in the third gradient between SSNHL (n = 30) and HC groups (n = 30) assessed by independent samples t-tests. (A) In the third gradient, the region-level differences of the principal gradient between two groups (two-sample t-test, P < 0.05, FDR uncorrected). The heat map scale bar displays the range of t-values for distinct brain regions, where the magnitude of these values reflects the degree of between-group differences in each region. Larger absolute values indicate greater between-group differences in that brain region. (B) Group differences in the second gradient values at the network level (two-sample t-test, FDR-uncorrected P < 0.05). Significant inter-group differences were observed in CON, VIS and SOM between the two groups. (C) Group differences in the TP that had significance differences (two-sample t-test, FDR-corrected P < 0.05). (D) Group differences in the PFC that had significance differences (two-sample t-test, FDR-corrected P < 0.05). TP, temporal pole; vPFC, ventral prefrontal cortex; CON, control; VIS, visual; SOM, somatomotor. *P < 0.05; **P < 0.01; ****P < 0.001.

Figure 7

Correlation between gradient values of ROIs and scale scores with linear regression analysis. (A) The primary gradient value of the TP in SSNHL patients (n = 30) showed a significant positive correlation with the SDMT scores (P = 0.034, r = 0.39). (B) The primary gradient value of the TP in SSNHL patients (n = 30) showed a significant positive correlation with the AVLT scores (P = 0.005, r = 0.50). (C) The primary gradient value of the TP in SSNHL patients (n = 30) showed a significant positive correlation with the AVLT_delay scores (P = 0.024, r = 0.41). (D) The primary gradient value of the vPFC in SSNHL patients (n = 30) showed a significant negative correlation with the HAMD scores (P = 0.013, r = −0.45). ROIs, regions of interest; TP, temporal pole; SSNHL, sudden sensorineural hearing loss; SDMT, Symbol Digit Modalities Test; AVLT, Auditory Verbal Learning Test; vPFC, ventral prefrontal cortex; HAMD, Hamilton Depression Scale.