Neuroscience 285 (2015) 333–342

CHANGES IN THE DEFAULT MODE NETWORKS OF INDIVIDUALS WITH LONG-TERM UNILATERAL SENSORINEURAL HEARING LOSS G.-Y. ZHANG, a,c* M. YANG, a B. LIU, a Z.-C. HUANG, b H. CHEN, b P.-P. ZHANG, b J. LI, a J.-Y. CHEN, a L.-J. LIU, d J. WANG d,e AND G.-J. TENG a

changes in the DMN have been found in the left than right long-term USNHL (RUSNHL). However, the neuropsychological tests did not show significant differences between the USNHL and the control. These findings suggest that longterm USNHL contributes to changes in the DMN, and these changes might affect cognitive abilities in patients with long-term USNHL. Left hearing loss affects the DMN more than the right hearing loss does. The fMRI measures might be more sensitive for observing cognitive changes in patients with hearing loss than clinical neuropsychological tests. This study provides some insights into the mechanisms of the association between hearing loss and cognitive function. Ó 2014 IBRO. Published by Elsevier Ltd. All rights reserved.

a Department of Radiology, Jiangsu Key Laboratory of Molecule Imaging and Functional Imaging, Medical School of Southeast University, 87 Dingjiaqiao Road, Nanjing 210009, China b Department of Otorhinolaryngology and Head-neck Surgery, ZhongDa Hospital, Medical School of Southeast University, 87 Dingjiaqiao Road, Nanjing 210009, China c

Department of Radiology, Taishan Medical University, Chang Cheng Road, Hi-Tech Development Zone, Taian 271016, Shandong Province, China d Department of Physiology and Pharmacology, Medical School of Southeast University, 87 Dingjiaqiao Road, Nanjing 210009, China e School of Human Communication Disorder, Dalhousie University, 1256 Barrington St, Halifax B3J1Y6, Canada

Key words: unilateral sensorineural hearing loss, default mode network, functional connectivity, nodal topological property, cognitive function.

Abstract—Hearing impairment contributes to cognitive dysfunction. Previous studies have found changes of functional connectivity in the default mode network (DMN) associated with cognitive processing in individuals with sensorineural hearing loss (SNHL). Whereas the changes in the DMN in patients with long-term unilateral SNHL (USNHL) is still not entirely clear. In this work, we analyzed resting-state functional magnetic resonance imaging (fMRI) data and neuropsychological test scores from normal hearing subjects (n = 11) and patients (n = 21) with long-term USNHL. Functional connectivity and nodal topological properties were computed for every brain region in the DMN. Analysis of covariance (ANCOVA) and post hoc analyses were conducted to identify differences between normal controls and patients for each measure. Results indicated that the left USNHL presented enhanced connectivity (p < 0.05, false discovery rate (FDR) corrected), and significant changes (p < 0.05, Bonferroni corrected) of the nodal topological properties in the DMN compared with the control. More

INTRODUCTION Hearing loss and cognitive dysfunction commonly occur in adults over the age of 60 (Gennis et al., 1991). Several investigations on bilateral hearing loss have found an association between hearing and cognitive dysfunction, these studies suggest that hearing impairment contributes to cognitive dysfunction (Tay et al., 2006; Lin, 2011; Lin et al., 2011, 2013). However, it remains unclear whether partial hearing deprivation would similarly affect cognitive processing in patients with unilateral sensorineural hearing loss (USNHL). The default mode network (DMN) (Raichle et al., 2001) is associated with cognitive processes—for example, emotional processing and self-referential mental activity (Lane et al., 1997; Gusnard et al., 2001), conflict monitoring (Kerns et al., 2004), memory retrieval (Wheeler et al., 2006) and cognitive control (Leech et al., 2011). Functional magnetic resonance imaging (fMRI) investigations have found enhanced functional connectivity (Husain et al., 2014; Wang et al., 2014) in the DMN in patients with sensorineural hearing loss (SNHL). Whereas task-based fMRI study reported an opposite result that children with USNHL displayed reduced deactivation of anterior and posterior regions of the DMN (Schmithorst et al., 2014). The former focuses on children, while the latter focuses on an adult population, which perhaps makes the difference. However, to update, the changes in the DMN in patients with long-term USNHL is still not entirely clear, and whether these

*Correspondence to: G.-Y. Zhang, Department of Radiology, Taishan Medical University, Chang Cheng Road, Hi-Tech Development Zone, Taian 271016, Shandong Province, China. Tel: +86-05386222723. E-mail address: [email protected] (G.-Y. Zhang). Abbreviations: ANCOVA, analysis of covariance; AVM, auditory verbal memory; AVMTR, auditory verbal memory test-delayed recall; BAs, Brodmann’s areas; BDS, backward digit spans; DMN, default mode network; DSST, digit symbol substitution test; FDR, false discovery rate; FDS, forward digit spans; fMRI, functional magnetic resonance imaging; HL, hearing level; LUSNHL, left long-term USNHL; MMSE, mini-mental state examination; MNI, Montreal Neurological Institute; MPFC, medial prefrontal cortex; ROI, region of interest; RUSNHL, right long-term USNHL; SCWA, Stroop color–word test A; SCWB, Stroop color–word test B; SCWC, Stroop color–word test C; SNHL, sensorineural hearing loss; SS, semantic similarity; TMTA, trail making test part A; TMTB, trail making test part B; USNHL, unilateral SNHL; VF, verbal fluency. http://dx.doi.org/10.1016/j.neuroscience.2014.11.034 0306-4522/Ó 2014 IBRO. Published by Elsevier Ltd. All rights reserved. 333

334

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

changes impact the cognitive abilities of patients is also unclear. Mathematical models for computational network analysis based on graph theory have been applied to brain connectivity, in which the brain network is modeled as a graph with the nodes representing different brain regions, and the edges representing interregional connectivity (Whitlow et al., 2011). The topological properties of the entire brain network reflect the entire performance of the network, whereas the nodal topological properties of the brain regions reflect local performance of the network. Graph theory network analyses have demonstrated that patients with central nervous systemic dysfunction showed changes in the topological properties of the brain network (He et al., 2009; Bai et al., 2012). A recent study on nodal topological properties found that more topological clustering was associated with impaired cognitive performance (Giessing et al., 2013). Therefore, we presume that long-term USNHL contributes to changes of functional connectivity in the DMN of patients, and further affects the nodal topological properties of the brain network. These changes impact cognitive processing. We expect that investigation on changes in the DMN can provide some insights into the mechanisms of the association between hearing loss and cognitive function.

EXPERIMENTAL PROCEDURES Subjects The protocol of this prospective study was approved by the institutional Ethics Committee of Southeast University, Nanjing, China, and all participants or their guardians signed informed consent forms prior to the experiment. A total of 32 subjects were recruited from Nanjing. Specifically, 11 patients (age, 27–60 years; mean = 47.2 ± 10.8 years; duration of the hearing loss, 2–50 years; mean = 14.9 ± 17.9 years; age at onset of unilateral hearing loss, 5–58 years, mean = 32.27 ± 19.42 years) had left long-term USNHL (LUSNHL) without tinnitus, 10 patients (age, 42–66 years; mean = 55.2 ± 6.8 years; duration of the hearing loss, 2–50 years; mean = 13.3 ± 14.6 years, age at onset of unilateral hearing loss, 16–55 years, mean = 41.90 ± 12.80 years) had right long-term USNHL (RUSNHL) without tinnitus, and 11 (age, 27–67 years; mean = 51.7 ± 12.4 years) were normal hearing volunteers. Hearing loss was acquired due to right acoustic neuroma in one patient, deafness was caused by infection in two patients, and hearing loss in others was unknown cause. All were right-handed and none of them had a history of medical, psychiatric disorders, strokes/cerebrovascular ischemia, head trauma/ fractures. The frequency-averaged speech-frequency thresholds (0.25, 0.5, 1.0 and 2 kHz) of every individual was measured by pure tone audiometry. For the USNHL groups, the average hearing level (HL) was at least 70 dB in the deaf ear and less than 25 dB in the better hearing ear. All unilaterally deaf subjects had post-lingual deafness. Eleven control subjects were healthy volunteers with pure tone average below 25 dB

HL for both ears. Cognitive abilities of participants were measured by the neuropsychological tests. No subjects were excluded because of excessive head movement (greater than 2.5 mm or 2.5°) during imaging. MR imaging All imaging was performed on a SIEMENS Verio 3-T scanner. Each subject underwent an 8.06-min scan during a conscious resting-state. Functional images were collected axially using an echo-planar imaging sequence sensitive to blood oxygen level dependent (BOLD) contrast. The following parameters were used: 2000/25 ms (repetition time/echo time), 36 slices, 64  64 (matrix size), 3.75  3.75  4 mm (voxel size), 1 mm (gap), 240  240 mm (field of view), 90° (flip angle). A high-resolution, three-dimensional T1-weighted structural image was also acquired for each subject using a magnetization prepared gradient echo sequence [repetition time (TR) = 1900 ms, echo time (TE) = 2.48 ms, flip angle = 9°, 176 slices with 0.975  0.975  1 mm voxels]. During the resting-state scanning, the light was switched off. The subjects wore earplugs and headphones to reduce noise, and were instructed to keep still with their eyes closed, as motionless as possible and not to think about anything in particular. Head movement was minimized by placing soft pads at the sides of the head. All subjects did not fall asleep, as confirmed by the subjects after completion of the experiment. Neuropsychological tests Mini-mental state examination (MMSE) and Wechsler adult intelligence test (Wechsler, 1997) were used to assess the cognitive abilities of the participants. Tests were administered in a quiet setting with minimal distractions. Experienced examiners who were accustomed to working with deaf patients verbally gave all instructions in a face-to-face manner. All of the neuropsychological tests were provided in Mandarin. Participants were tested on (a) MMSE, (b) digit symbol substitution test (DSST), (c) forward digit spans (FDS) and backward digit spans (BDS), (d) verbal fluency (VF), (e) semantic similarity (SS), (f) trail making test part A (TMTA), (g) trail making test part B (TMTB), (h) Stroop color–word test A (SCWA), (i) Stroop color–word test B (SCWB), (j) Stroop color– word test C (SCWC), (k) auditory verbal memory (AVM), and (l) auditory verbal memory test-delayed recall (AVMTR). MMSE was performed for likely cognitive impairment, with test components covering concentration, language, and memory (Tay et al., 2006). DSST performance recruited different interrelated abilities such as perceptual speed, motor speed, response selection, and shifting of attention, as well as working memory (Fratiglioni et al., 2000). FDS was thought to measure the storage and maintenance components of working memory, and BDS was a complex task that relies heavily on working memory processing because it requires information storage as well as concurrent processing essential

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

to mentally reordering the information (Horovitz et al., 2009). VF test was one of the easiest methods in the neuropsychological evaluation of the frontal and temporal lobe function (Wysokinski et al., 2010). SS test has been proposed to measure the abilities of semantic processing and executive control (Allen et al., 2012). TMTA was widely used to assess basic executive abilities, such as motor speed (Tamez et al., 2011). SCWA, SCWB, and SCWC tests were implemented for examining attention and response inhibition (Kenworthy et al., 1990). Data processing Data sets were preprocessed by using DPARSF V2.0 (http://www.restfmri.net/forum/DPARSF). The first 10 time points of each subject were discarded to allow for the stabilization of the MR signals. The remaining functional scans were corrected for within-scan acquisition time differences between slices and realigned to the first volume to correct for within-run head motions by using a six-parameter affine transformation. Images were registered with structural images and normalized to the standard brain template from the Montreal Neurological Institute (MNI 152) by using nonlinear registration. The normalized fMRI data were entered into the functional connectivity software CONN toolbox (http:// www.nitrc.org/proects/conn), which executed the following steps: first, the normalized data were temporally filtered by using a band-pass filtering (0.01– 0.1 Hz) to reduce the effect of high-frequency physiologic noise; subsequently, white matter, gray matter, cerebrospinal fluid, realignment parameters, age and educational level were taken as confounders, following the implemented CompCor strategy (Behzadi et al., 2007). The CONN toolbox performed region of interest (ROI)-based analysis by grouping voxels into ROIs on the basis of Brodmann’s areas (BAs). Bi-variate correlations were calculated between each pair of ROIs as reflections of connections. Correlation coefficients were transformed into z-scores by using Fisher z transforms to improve the normality (Slobounov et al., 2011). All BAs were imported as possible connections for each selected ROI source. A one-way ANOVA and post hoc analyses were executed to identify changes in functional connectivity in each brain region of the DMN. In this study, the dependent factor was z-scores, and the independent factor was group (level). Three levels were used for each factor. We constructed the entire brain functional network (graph) with N nodes and K edges by using CONN software based on the z-scores. The nodes of the network are mainly labeled as the number of BAs. For those regions that are not well indicated in the BA system, anatomical names are given as ROIs. Graph theory metrics (i.e., the shortest path length, clustering coefficient, and nodal degree) were computed by using the CONN functional connectivity toolbox. The nodal topological properties observed in this study involve the shortest path length, nodal degree, and clustering coefficient. For the nodal topological analysis, the brain is modeled as a network (graph) with N nodes

335

representing brain regions and K edges representing interregional connectivity (Whitlow et al., 2011). The shortest path length Li of a node i is a measure of the capacity of parallel information transfer from one node to the rest of the network and is estimated by the average over all minimum number of edges in the paths from node i to all other nodes in the network (Watts and Strogatz, 1998). A smaller value of Li indicates greater efficiency of information transfer from one node to the others in the network. The nodal degree of a node is measured by the number of connections to that node and reflects the extent of interconnectivity with other nodes or the extent of functional reorganization of this node. The clustering coefficient Ci of a node i is defined as the number of existing connections in the neighborhood of this node divided by all its possible connections. This parameter measures the extent of local density or the cliquishness of a typical neighborhood (a local property) and quantifies the local efficiency of information transfer of a network (Watts and Strogatz, 1998). We threshold the graph with sparsity S, which is defined as K divided by the maximum possible number of edges N(N 1)/2. Setting a sparsityspecific threshold ensures that the estimated betweengroup difference can purely reflect the alterations in the topological organization (He et al., 2008). We compared the topological parameters over a wide range of sparsity (5% 6 S 6 50%) across the three groups. VBM analysis Gray matter volume of the primary auditory cortex (BAs 41 and 42) was assessed by a voxel-based morphometry (Wright et al., 1995, 1999; Ashburner and Friston, 2000), using the SPM8 software and in-house C++ codes. The following procedures were applied: (1) a nonlinear transformation T was obtained by using the nonlinear registration from the T1-weighted structural MRI to the ICBM152 T1 template in the MNI space. (2) The inverse transformation T 1 was applied to the BA template in the MNI space and resulted in a subject-specific BA mask in the structural MRI native space. (3) The structural MRI was segmented into gray matter and white matter, and gray matter volume of the primary auditory cortex (BAs 41and 42) of each subject was calculated based on the subject-specific BA mask. Statistical analyses An analysis of covariance (ANCOVA) was performed to remove the possible confounding effects of demographic variables (age, educational level), and explore group differences in the nodal topological properties (shortest path length, nodal degree, and clustering coefficient) and the neuropsychological test scores. In this study, the dependent variables were the nodal topological properties and the neuropsychological test scores, the independent variables were groups, and the covariant variables were the demographic variables. In addition, paired t-tests were executed to explore the differences of gray matter volume between the left and right primary auditory cortex in the USNHL groups. Statistical analyses were conducted with software (SPSS, version

336

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

19.0). Bonferroni correction was used for multiple comparisons. Three comparisons were done in every analysis, and a p value of less than .0167 (.05/ 3 = .0167) was considered to indicate a significant difference for any single analysis.

RESULTS Demography and hearing threshold test No significant differences in gender and age are noted across the three groups (p > 0.2). However, an ANOVA reveals a significant difference for educational level among the three groups [LUSNHL, 9–14 years, mean = 12.6 ± 1.8 years; RUSNHL, 9–14 years, mean = 12.4 ± 2.2 years; normal, 12–18 years, mean = 15.1 ± 2.1 years; F = 5.79(2, 39), p = 0.008]. Post hoc analysis revealed that the USNHL presented significantly lower educational level compared with the control (p < 0.025). No significant difference is found between the two USNHL groups for frequency averaged hearing thresholds and duration of hearing loss (p > 0.506). In addition, two sample t-test showed no significant difference of the age at onset of unilateral hearing loss between the L and RUSNHL (T(2,19) = 1.326, p = 0.201). Neuropsychological test results The neuropsychological characterizations for all three groups are summarized in Table 1. A lower score means a better performance in TMTA, TMTB, SCWA, SCWB, and SCWC tests. By contrast, a lower score indicates a worse performance in MMSE, DSST, VF, AVM, AVMTR, FDS, BDS, and SS tests. However, no significant difference was found in all neuropsychological tests across the three groups.

Cross-group differences of functional connectivity in the DMN The DMN as described in previous studies (Raichle et al., 2001; Damoiseaux et al., 2006; Sorg et al., 2007) consists of the superior medial frontal (BA 8/9/10), medial prefrontal cortex (MPFC), anterior cingulate (BA 32/24), posterior cingulate/precuneus (BA 23/31), middle and inferior temporal (BA 21/20), and the inferior parietal region. The difference of functional connectivity in each seeding region in the DMN across the three groups was identified using a one-way ANOVA and post hoc analysis. BAs and other ROIs defined by anatomical labels in the DMN were analyzed in this current study. However, only those regions showing significant changes (p < 0.05, false discovery rate (FDR) corrected, and 100 ROIs were used in the FDR correction) are described as follows. In Fig. 1, compared with the control, the LUSNHL shows stronger connectivity for the left posterior cingulate (BA 23L) with the primary somatosensory cortex (BAs 1L, 3R), right primary motor cortex (BA 4R), premotor cortex (BA 6), superior temporal gyrus (BA 22), left primary auditory cortex (BA 41L), left somatosensory association cortex (BA 5L), and the left postcentral gyrus (BA 43L) (Fig. 1A); precuneus with the right primary somatosensory cortex (BA 3R), primary motor cortex (BA 4), somatosensory association cortex (BA 5), right premotor cortex (BA 6R), left anterior cingulate (BA 24L), left primary auditory cortex (BA 41L/42L), left postcentral gyrus (BA 43L), left secondary visual cortex (BA 18L), right associative visual cortex (BA 19R), and the right fusiform gyrus (BA 37R) (Fig. 1B); the left inferior temporal gyrus (BA 20L) with the primary somatosensory cortex (BAs 1L, 3), primary motor cortex (BA 4), left somatosensory association cortex (BA 5L),

Table 1. Summary of behavior tests Tests

L (11)

R (10)

N (11)

FLev

PLev *

F2,29

P anova

*

P post hoc LR: 0.410** LN: 0.267** RN: 0.990**

TMTA

100.5 ± 1.34

99.79 ± 0.85

99.69 ± 0.72

3.976

0.03

TMTB SCWA SCWB SCWC MMSE

99.97 ± 0.58 100.6 ± 3.54 99.97 ± 1.80 99.89 ± 1.22 95.0 ± 14.1

100.0 ± 0.58 99.99 ± 2.31 99.76 ± 1.79 100.0 ± 0.74 102.2 ± 5.7

100.0 ± 0.45 99.40 ± 2.82 100.3 ± 1.72 100.1 ± 0.75 103.0 ± 8.3

0.599 0.664 0.164 1.249 7.301*

0.556 0.522 0.849 0.302 0.003*

0.026 0.452 0.207 0.120

DSST VF FDS BDS AVM AVMTR SS

99.70 ± 1.59 99.84 ± 3.27 99.73 ± 13.4 96.14 ± 6.41 99.96 ± 0.48 98.27 ± 8.14 97.50 ± 5.90

99.97 ± 1.94 99.68 ± 2.60 103.8 ± 10.8 102.1 ± 9.88 99.96 ± 0.50 101.6 ± 8.64 100.8 ± 5.81

100.3 ± 1.44 100.5 ± 2.33 96.78 ± 9.06 101.97 ± 12 100.1 ± 0.4 100.3 ± 6.81 101.8 ± 2.87

0.051 0.263 0.941 1.093 0.182 0.408 2.147

0.951 0.771 0.402 0.349 0.834 0.669 0.135

0.392 0.233 1.036 1.326 0.255 0.481 2.172

0.975 0.640 0.814 0.887 LR: 0.337** LN: 0.335** RN: 0.992** 0.679 0.794 0.368 0.281 0.776 0.623 0.132

Data are in terms of mean ± standard deviation. N, normal controls; L, left unilateral sensorineural hearing loss (LUSNHL); R, right unilateral sensorineural hearing loss (RUSNHL). TMTA, trail making test part A; TMTB, trail making test part B; SCWA, Stroop color–word test A; SCWB, Stroop color–word test B; SCWC, Stroop color–word test C; MMSE, mini-mental state examination; DSST, digit symbol substitution test; VF, verbal fluency; FDS, forward digit spans; BDS, backward digit spans; AVM, auditory verbal memory; AVMTR, auditory verbal memory test-delayed recall. * Homogeneity test (Levene’s test). ** Tamhane’s T2 test.

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

Fig. 1. Diagrams of functional connectivity across the three groups. The diagrams were developed by CONN toolbox based upon p < 0.05, FDR corrected in post hoc t-tests between the USNHL and the control. (A–G) are between the LUSNHL and the control. (H–I) are between the L and RUSNHL. The seeds are represented by white circles or spheres, and the number in the circle indicates the BA-based ROIs. The red line indicates an increase in functional connectivity. Arrow-widths represent T-values. LLP, left inferior parietal lobule (MNI, -47, -74, 35); LIPL, left inferior parietal lobule (MNI, -42, -72, 38); MPFC, medical prefrontal cortex. See CONN toolbox manual for details. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

premotor cortex (BA 6), left anterior cingulate (BA 24L/ 32L), left primary auditory cortex (BA 41L/42L), left postcentral gyrus (BA 43L), the right subgenual cortex (BA 25R) and left piriform cortex (BA 27L) (Fig. 1C); the MPFC with the left primary motor cortex (BA 4L) (Fig. 1D); the left superior medial frontal (BA 8L) with the left primary auditory cortex (BA 42L) (Fig. 1E); the left inferior parietal lobule [LIPL (MNI, -42, -72, 38)] with the left anterior cingulated (BA 24L), the left primary auditory cortex (BA 42L) and postcentral gyrus (BA 43L) (Fig. 1F); the left anterior cingulated (BA 24L) with the left inferior parietal lobule [LLP (MNI, -46, -74, 35), LIPL (MNI, -42, -72, 38)] and retrosplenial cingulate (BA 29L) (Fig. 1G). Compared with the RUSNHL, the LUSNHL presents only stronger connectivity between the left inferior temporal gyrus (BA 20L) and primary motor cortex (BA 4L) (Fig. 1H); the right superior medial frontal (BA 8R) and left piriform cortex (BA 27L) (Fig. 1I). No significant differences in connectivity are seen between the RUSNHL and the control for any brain regions in the DMN. In addition, we also explored the changes of functional connectivity in the auditory and attention networks.

337

Significant changes (p < 0.005, uncorrected) are described as follows. In Fig. 2, compared with the control, the LUSNHL shows stronger connectivity for the left primary auditory cortex (BA 41L) with the left primary somatosensory cortex (BA 3L), left primary motor cortex (BA 4L), left superior medial frontal (BA 8L), left inferior temporal gyrus (BA 20L), posterior cingulated cortex (PCC), and the left angular gyrus (BA 39L) (Fig. 2A); the left primary auditory cortex (BA 42L) with the left superior medial frontal (BAs 8L, 9L), left inferior temporal gyrus (BA 20L), posterior cingulated cortex (PCC), left somatosensory association cortex (BA 7L), and the left inferior parietal lobule [LIPL (MNI, -42, -72, 38)] (Fig. 2B). Compared with the control, the RUSNHL presents enhanced functional connectivity between the right primary auditory cortex (BAs 41R, 42R) and the left primary visual cortex (BA 17L) (Fig. 2C), and weak connectivity between the right inferior frontal cortex (BA 44R) in the attention network and the right anterior cingulate cortex (BA 32R) (Fig. 2D). Compared with the RUSNHL, the LUSNHL shows stronger connectivity for the left primary auditory cortex (BA 41L) with the left superior medial frontal (BA 8L), left inferior temporal gyrus (BA 20L), and the left superior frontal gyrus (LSFG) (Fig. 2); the left primary auditory cortex (BA 42L) with the left superior medial frontal (BA 10L) (Fig. 2F). Altered nodal topological properties in the DMN in patients We performed an ANCOVA on the nodal topological properties for every brain region in the DMN. However, only those areas that presented significant main effects among the three groups were described below. An ANCOVA revealed that significant main effects were observed for several areas on the magnitudes of the shortest path lengths, the left inferior temporal gyrus (BA 20L): F2,29 = 3.928, p = .031; the left posterior cingulate (BA 31L): F2,29 = 5.533, p = .009; the left posterior cingulate (BA 23L): F2,29 = 7.636, p = .002; precuneus: F2,29 = 8.039, p = .002; and nodal degrees, BA 20L: F2,29 = 3.653, p = .038; BA 31L: F2,29 = 6.231, p = .006; BA 23L: F2,29 = 8.189, p = .002; precuneus: F2,29 = 7.380, p = .003. Post hoc analyses revealed that the left inferior temporal gyrus (BA 20L), and the left posterior cingulate/precuneus (BA 23L/31L) exhibited a significant decrease for the shortest path lengths and increase for nodal degree in the LUSNHL compared with the normal control (Fig. 3A–H). The RUSNHL also showed a significant decrease for the nodal degree in the left posterior cingulate (BA 23L/31L) compared with the LUSNHL (Fig. 3D, G). Comparisons of nodal topological properties as function of sparsity over a wide range (5–50%) across the three groups are shown in Fig. 4. We also performed a voxel-based morphometry analysis (Wright et al., 1995, 1999; Ashburner and Friston, 2000) to explore whether there was actually an asymmetry presented in the primary auditory cortex (BAs 41 and 42). Paired t-test revealed a significant difference of gray matter volume between the left and right

338

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

Fig. 2. Diagrams of functional connectivity across the three groups. The diagrams were developed by CONN toolbox based upon p < 0.005, uncorrected in post hoc t-tests between the USNHL and the control. (A, B) are between the LUSNHL and the control. (C, D) are between the RUSNHL and the control. (E, F) are between the L and RUSNHL. The seeds are represented by white spheres, and the number beside the sphere indicates the BA-based ROIs. The red line indicates an increase in functional connectivity, and the blue line indicates a decrease in functional connectivity. Arrow-widths represent T-values. LIPL, left inferior parietal lobule (MNI, -42, -72, 38); LSFG, left superior frontal gyrus; PCC, posterior cingulated cortex. See CONN toolbox manual for details. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

primary auditory cortical regions in the LUSNHL [left = 683.64 ± 61.76 (mm3), right = 653.82 ± 71.26 (mm3), T(2,10) = 2.85, p = 0.017, uncorrected], and presented a similar nonsignificant trend in the RUSNHL [left = 620.40 ± 50.45 (mm3), right = 617.10 ± 38.16 (mm3), T(2,9) = 0.333, p = 0.747, uncorrected]. In addition, Pearson’s correlation analysis showed no significant relationship between this asymmetry of gray matter volume and the duration of the disease, or time of onset (R > 0.116, p > 0.615, uncorrected).

DISCUSSION We investigated the changes in the DMN induced by hearing loss. Our main findings are as follows. The left USNHL presented enhanced connectivity (p < 0.05, FDR corrected), and significant changes (p < 0.05, Bonferroni corrected) of the nodal topological properties in several brain regions involved in cognitive processing, such as the left posterior cingulate/precuneus (BA 31L/ 23L) and inferior temporal (BA 20L) compared with the control. The LUSNHL showed more changes in the DMN compared with the RUSNHL. Hearing loss-related sensory decline increases listening effort, which has often been associated with increased activity in some brain regions related with cognitive processing (Peelle et al., 2011). In this study, several areas in the DMN of the LUSNHL showed enhanced functional connectivity. These areas included the left anterior cingulate (BA 24L) that has been found to be associated with emotional processing, conflict monitoring (Lane et al., 1997; Kerns et al., 2004); left superior medial frontal (BA 8L/9L/10L) and MPFC, which are responsible for self-referential mental activity (Gusnard et al., 2001); the left posterior cingulated/precuneus (BA 23L/31L), which is involved in cognitive control (Leech

et al., 2011), and the left inferior temporal (BA 20L) that is responsible for memory retrieval (Wheeler et al., 2006). Our finding is consistent with the previous report (Husain et al., 2014; Wang et al., 2014) that the patients with hearing loss presented increased functional connectivity in the DMN. Reduced DMN connectivity at rest is associated with cognitive impairment (Wang et al., 2006; Kennedy and Courchesne, 2008). Therefore, the enhanced connectivity in the DMN of the LUSNHL might be a compensation for the decline of cognition. Decline in hearing ability contributes to loss of gray matter volume in the primary auditory cortex (Peelle et al., 2011; Lin et al., 2014), which impairs the capacity of the auditory cortical areas to respond to sound stimulation. These changes may affect cognitive processing of brain functional networks. To limit the consequences of neurological damage and to help maintain cognitive abilities (Hawellek et al., 2011), a plastic reorganization occurs to compensate for the impairment of cognition induced by long-term hearing loss. Such functional reorganization is manifested as enhanced functional connectivity in some brain regions in the DMN. Cochlear implantation in patients with hearing loss yields a partial recovery of auditory functions and speech comprehension, which is expressed as an increment of resting-state activity in the auditory and visual cortex, and the magnitude of activity is positively correlated with the ability of word recognition (Strelnikov et al., 2010, 2013). The current data reveal enhanced connectivity between the left primary auditory cortex (BA 41L/42L) and the DMN, as well as between the visual cortex (BAs 18L) and the DMN in the LUSNHL. These results indicate an improvement of auditory and cognitive abilities in the left hemisphere in the LUSNHL, which compensate the decline of auditory and cognitive abilities in the right hemisphere induced by degraded auditory signals.

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

Fig. 3. Cross-group comparisons of nodal topological properties for BAs 20L, 31L, 23L, and precuneus. (A–H) show the bar graphs of the nodal property at the sparsity threshold of 20% for corresponding brain regions and the results of post hoc pairwise comparisons. Each significant difference (p < 0.05, Bonferroni corrected) in any pair of two groups is indicated by the p values. Error bars: ±2SEs (standard errors).

Specific auditory areas in early deaf subjects can be recruited for visual (Nishimura et al., 1999; Bavelier et al., 2000; Petitto et al., 2000; Finney et al., 2001) processing. However, this specific and localized cross-modal reorganization in patients with hearing loss may result in a functional disintegration of auditory cortical areas, which causes deficits in some cognitive functions (Kral and Sharma, 2012). Cochlear-implant users showed activation in the right auditory cortex for visual processing, which was negatively related to speech recognition ability with a cochlear implant (Sandmann et al., 2012). Our current data showed an enhanced connectivity between the right primary auditory (BAs 41R, 42R) and the left primary visual (BA 17L) in the RUSNHL. This result implies that right long-term USNHL contributes to cross-modal functional reorganization that might affect cognitive abilities of patients with right hearing loss. We also noted that stronger connectivity was related to decease in the shortest path length or increase in nodal degree in the left inferior temporal (BA 20L) and the left posterior cingulated/precuneus (BA 23L/31L). Previous studies have reported that poorer cognitive

339

Fig. 4. Comparisons of nodal topological properties across the three groups for BA 20L, 31L, 23L, and precuneus. A–H show multiline graphs of the nodal properties as the function of sparsity over a wide range (5–50%). The asterisks marked the points at which a significant difference (p < 0.05) is seen among the three groups. Error bars: ±2 SEs.

performance was correlated with longer characteristic path length (He et al., 2008; Bai et al., 2012). Therefore, shorter characteristic path length in the DMN in patients with long-term USNHL might be a compensatory for cognitive decline induced by hearing loss. The LUSNHL showed more changes in the primary auditory cortex and the DMN compared with the RUSNHL. Moreover, changes focused on the left hemisphere. One possible reason is that the LUSNHL presents greater gray matter volume in the left primary auditory cortex (BA 41 and 42L) than the right. The structural asymmetry contributes to more changes of plasticity in the left primary auditory cortex, and these changes further impact the DMN. In addition, the difference of the age at onset of hearing loss might also impact the plasticity of the cerebral cortex. Burton (Burton et al., 2012) studied the whole auditory cortical field and found the primary auditory cortex showed left

340

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

hemisphere dominance with right or left ear stimulation, but other auditory cortical fields had greater reorganization in the right hemisphere in the LUSNHL. The inconsistent results might be caused by the difference of the age at onset of hearing loss. The mean age is more than 32 years in our study, and less than 25 years in previous study (Burton et al., 2012). Additionally, Burton studied activation magnitudes in the whole auditory cortical field using task-based fMRI (Burton et al., 2012), and our investigation used resting-state fMRI. Different study modes might also caused the inconsistent results. Other study also found more changes of resting-state functional connectivity in the left unilateral hearing loss patients than the right ones (Wang et al., 2014). In most studies of USNHL, it has been found either that there is no difference between ears (Scheffler et al., 1998; Schmithorst et al., 2005; Burton et al., 2012), or that right ear deafness affects academic and language performance more than left ear (Bess and Tharpe, 1986a,b; Hartvig Jensen et al., 1989; Schmithorst et al., 2014). The findings of this current study indicate reorganization of the DMN in left > right ear deafness. One possible reason is that previous studies were mainly focused on children whose mean age was less than 13 years, and the language acquisition period in humans lasts about 13 years (Komarova and Nowak, 2001). Worse language and receptive speech outcomes were observed in the children with RUSNHL compared to the children with LUSNHL, since the auditory projection is contralaterally dominated (Hugdahl, 2000), the auditory signal from the left ear is predominantly processed in the right hemisphere, and partial signal is transferred to the language processing centers, which are in the left hemisphere for most individuals (Schmithorst et al., 2014). In this current study, all unilaterally deaf subjects have post-lingual deafness, and their mean age is more than 47 years. Investigations on bilateral hearing loss have indicated that hearing impairment contributed to cognitive dysfunction (Tay et al., 2006; Lin, 2011; Lin et al., 2011, 2013). In our study, although the USNHL presented worse performance than the control in most of behavior tests such as TMTA, SCWA, DSST, MMSE, VF, AVM, and SS, in which significant difference was not reached. One possible reason is that partial hearing loss-related functional reorganization increases cognitive processing abilities of some brains. As a result, the cortical reorganization limits the decline of cognitive abilities in patients with long-term USNHL. Another explanation is that aging is associated with changes in cognitive function accompanied with an increased shortest path length during the performance of cognitive tasks (Wang et al., 2010; Peelle et al., 2011). The significant differences of nodal topological properties in more brain regions of the DMN might be found with aging. These changes contribute to the decline of cognitive function. Previous study indicated that greater amounts of hearing loss were associated with a progressively higher risk of dementia and cognitive dysfunction in older adults (Uhlmann et al., 1989). The mean age of participants in our study is less than 56, whereas previous studies were mainly focused on adults over age of 60. Hence,

the difference of age may be one reason that causes inconsistent results. In addition, we also explored the dorsal and ventral attention network, and the auditory network. The right inferior frontal gyrus (BA 44R) in the attention network presented weak functional connectivity with the DMN in the RUSNHL. The left primary auditory cortex displayed enhanced functional connectivity with the DMN in the LUSNHL. However, no significant change was observed within the auditory network. In this study, the effect of education level has been eliminated by using a component based noise correction method (CompCor) (Behzadi et al., 2007) or an ANCOVA to ensure all observed effects are caused by the hearing loss and not by the significant difference in education level. In a word, these findings suggest that long-term USNHL contributes to changes in the DMN, and these changes might affect cognitive abilities in patients with long-term USNHL. However, in our study, behavior tests did not show significant differences between the USNHL and the control. These results suggest that the fMRI measures might be more sensitive for observing cognitive changes in patients with hearing loss than clinical neuropsychological tests. This study provides some insights into the mechanisms of the association between hearing loss and cognitive function. Acknowledgments—This research was supported by the National Natural Science Foundation of China (grant no. 30970808).

REFERENCES Allen CM, Martin RC, Martin N (2012) Relations between short-term memory deficits, semantic processing, and executive function. Aphasiology 26:428–461. Ashburner J, Friston KJ (2000) Voxel-based morphometry–the methods. Neuroimage 11:805–821. Bai F, Shu N, Yuan Y, Shi Y, Yu H, Wu D, Wang J, Xia M, He Y, Zhang Z (2012) Topologically convergent and divergent structural connectivity patterns between patients with remitted geriatric depression and amnestic mild cognitive impairment. J Neurosci 32:4307–4318. Bavelier D, Tomann A, Hutton C, Mitchell T, Corina D, Liu G, Neville H (2000) Visual attention to the periphery is enhanced in congenitally deaf individuals. J Neurosci 20:RC93. Behzadi Y, Restom K, Liau J, Liu TT (2007) A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage 37:90–101. Bess FH, Tharpe AM (1986a) Case history data on unilaterally hearing-impaired children. Ear Hear 7:14–19. Bess FH, Tharpe AM (1986b) An introduction to unilateral sensorineural hearing loss in children. Ear Hear 7:3–13. Burton H, Firszt JB, Holden T, Agato A, Uchanski RM (2012) Activation lateralization in human core, belt, and parabelt auditory fields with unilateral deafness compared to normal hearing. Brain Res 1454:33–47. Damoiseaux JS, Rombouts SA, Barkhof F, Scheltens P, Stam CJ, Smith SM, Beckmann CF (2006) Consistent resting-state networks across healthy subjects. Proc Natl Acad Sci U S A 103:13848–13853. Finney EM, Fine I, Dobkins KR (2001) Visual stimuli activate auditory cortex in the deaf. Nat Neurosci 4:1171–1173.

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342 Fratiglioni L, Wang HX, Ericsson K, Maytan M, Winblad B (2000) Influence of social network on occurrence of dementia: a community-based longitudinal study. Lancet 355:1315–1319. Gennis V, Garry PJ, Haaland KY, Yeo RA, Goodwin JS (1991) Hearing and cognition in the elderly. New findings and a review of the literature. Arch Intern Med 151:2259–2264. Giessing C, Thiel CM, Alexander-Bloch AF, Patel AX, Bullmore ET (2013) Human brain functional network changes associated with enhanced and impaired attentional task performance. J Neurosci 33:5903–5914. Gusnard DA, Akbudak E, Shulman GL, Raichle ME (2001) Medial prefrontal cortex and self-referential mental activity: relation to a default mode of brain function. Proc Natl Acad Sci U S A 98:4259–4264. Hartvig Jensen J, Johansen PA, Borre S (1989) Unilateral sensorineural hearing loss in children and auditory performance with respect to right/left ear differences. Br J Audiol 23: 207–213. Hawellek DJ, Hipp JF, Lewis CM, Corbetta M, Engel AK (2011) Increased functional connectivity indicates the severity of cognitive impairment in multiple sclerosis. Proc Natl Acad Sci U S A 108:19066–19071. He Y, Chen Z, Evans A (2008) Structural insights into aberrant topological patterns of large-scale cortical networks in Alzheimer’s disease. J Neurosci 28:4756–4766. He Y, Chen Z, Gong G, Evans A (2009) Neuronal networks in Alzheimer’s disease. Neuroscientist 15:333–350. Horovitz SG, Braun AR, Carr WS, Picchioni D, Balkin TJ, Fukunaga M, Duyn JH (2009) Decoupling of the brain’s default mode network during deep sleep. Proc Natl Acad Sci U S A 106:11376–11381. Hugdahl K (2000) Lateralization of cognitive processes in the brain. Acta Psychol 105:211–235. Husain FT, Carpenter-Thompson JR, Schmidt SA (2014) The effect of mild-to-moderate hearing loss on auditory and emotion processing networks. Front Syst Neurosci 8:10. Kennedy DP, Courchesne E (2008) The intrinsic functional organization of the brain is altered in autism. Neuroimage 39:1877–1885. Kenworthy OT, Klee T, Tharpe AM (1990) Speech recognition ability of children with unilateral sensorineural hearing loss as a function of amplification, speech stimuli and listening condition. Ear Hear 11:264–270. Kerns JG, Cohen JD, MacDonald 3rd AW, Cho RY, Stenger VA, Carter CS (2004) Anterior cingulate conflict monitoring and adjustments in control. Science 303:1023–1026. Komarova NL, Nowak MA (2001) Natural selection of the critical period for language acquisition. Proc Biol Sci 268: 1189–1196. Kral A, Sharma A (2012) Developmental neuroplasticity after cochlear implantation. Trends Neurosci 35:111–122. Lane RD, Fink GR, Chau PM, Dolan RJ (1997) Neural activation during selective attention to subjective emotional responses. Neuroreport 8:3969–3972. Leech R, Kamourieh S, Beckmann CF, Sharp DJ (2011) Fractionating the default mode network: distinct contributions of the ventral and dorsal posterior cingulate cortex to cognitive control. J Neurosci 31:3217–3224. Lin FR (2011) Hearing loss and cognition among older adults in the United States. J Gerontol A Biol Sci Med Sci 66:1131–1136. Lin FR, Ferrucci L, An Y, Goh JO, Doshi J, Metter EJ, Davatzikos C, Kraut MA, Resnick SM (2014) Association of hearing impairment with brain volume changes in older adults. Neuroimage 90C:84–92. Lin FR, Ferrucci L, Metter EJ, An Y, Zonderman AB, Resnick SM (2011) Hearing loss and cognition in the Baltimore longitudinal study of aging. Neuropsychology 25:763–770. Lin FR, Yaffe K, Xia J, Xue QL, Harris TB, Purchase-Helzner E, Satterfield S, Ayonayon HN, Ferrucci L, Simonsick EM, Health ABCSG (2013) Hearing loss and cognitive decline in older adults. JAMA Int Med 173:293–299.

341

Nishimura H, Hashikawa K, Doi K, Iwaki T, Watanabe Y, Kusuoka H, Nishimura T, Kubo T (1999) Sign language ‘heard’ in the auditory cortex. Nature 397:116. Peelle JE, Troiani V, Grossman M, Wingfield A (2011) Hearing loss in older adults affects neural systems supporting speech comprehension. J Neurosci 31:12638–12643. Petitto LA, Zatorre RJ, Gauna K, Nikelski EJ, Dostie D, Evans AC (2000) Speech-like cerebral activity in profoundly deaf people processing signed languages: implications for the neural basis of human language. Proc Natl Acad Sci U S A 97:13961–13966. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard DA, Shulman GL (2001) A default mode of brain function. Proc Natl Acad Sci U S A 98:676–682. Sandmann P, Dillier N, Eichele T, Meyer M, Kegel A, Pascual-Marqui RD, Marcar VL, Jancke L, Debener S (2012) Visual activation of auditory cortex reflects maladaptive plasticity in cochlear implant users. Brain 135:555–568. Scheffler K, Bilecen D, Schmid N, Tschopp K, Seelig J (1998) Auditory cortical responses in hearing subjects and unilateral deaf patients as detected by functional magnetic resonance imaging. Cereb Cortex 8:156–163. Schmithorst VJ, Holland SK, Ret J, Duggins A, Arjmand E, Greinwald J (2005) Cortical reorganization in children with unilateral sensorineural hearing loss. Neuroreport 16:463–467. Schmithorst VJ, Plante E, Holland S (2014) Unilateral deafness in children affects development of multi-modal modulation and default mode networks. Front Hum Neurosci 8:164. Slobounov SM, Gay M, Zhang K, Johnson B, Pennell D, Sebastianelli W, Horovitz S, Hallett M (2011) Alteration of brain functional network at rest and in response to YMCA physical stress test in concussed athletes: RsFMRI study. Neuroimage 55:1716–1727. Sorg C, Riedl V, Muhlau M, Calhoun VD, Eichele T, Laer L, Drzezga A, Forstl H, Kurz A, Zimmer C, Wohlschlager AM (2007) Selective changes of resting-state networks in individuals at risk for Alzheimer’s disease. Proc Natl Acad Sci U S A 104:18760–18765. Strelnikov K, Rouger J, Demonet JF, Lagleyre S, Fraysse B, Deguine O, Barone P (2010) Does brain activity at rest reflect adaptive strategies? Evidence from speech processing after cochlear implantation. Cereb Cortex 20:1217–1222. Strelnikov K, Rouger J, Demonet JF, Lagleyre S, Fraysse B, Deguine O, Barone P (2013) Visual activity predicts auditory recovery from deafness after adult cochlear implantation. Brain 136:3682–3695. Tamez E, Myerson J, Morris L, White DA, Baum C, Connor LT (2011) Assessing executive abilities following acute stroke with the trail making test and digit span. Behav Neurol 24:177–185. Tay T, Wang JJ, Kifley A, Lindley R, Newall P, Mitchell P (2006) Sensory and cognitive association in older persons: findings from an older Australian population. Gerontology 52:386–394. Uhlmann RF, Larson EB, Rees TS, Koepsell TD, Duckert LG (1989) Relationship of hearing impairment to dementia and cognitive dysfunction in older adults. JAMA 261:1916–1919. Wang L, Li Y, Metzak P, He Y, Woodward TS (2010) Age-related changes in topological patterns of large-scale brain functional networks during memory encoding and recognition. Neuroimage 50:862–872. Wang L, Zang Y, He Y, Liang M, Zhang X, Tian L, Wu T, Jiang T, Li K (2006) Changes in hippocampal connectivity in the early stages of Alzheimer’s disease: evidence from resting state fMRI. Neuroimage 31:496–504. Wang X, Fan Y, Zhao F, Wang Z, Ge J, Zhang K, Gao Z, Gao JH, Yang Y, Fan J, Zou Q, Liu P (2014) Altered regional and circuit resting-state activity associated with unilateral hearing loss. PLoS One 9:e96126. Watts DJ, Strogatz SH (1998) Collective dynamics of ‘small-world’ networks. Nature 393:440–442. Wechsler D (1997) Wechsler Adult Intelligence Scale. 3. New York: Psychological Corporation. Wheeler ME, Shulman GL, Buckner RL, Miezin FM, Velanova K, Petersen SE (2006) Evidence for separate perceptual reactivation and search processes during remembering. Cereb Cortex 16:949–959.

342

G.-Y. Zhang et al. / Neuroscience 285 (2015) 333–342

Whitlow CT, Casanova R, Maldjian JA (2011) Effect of resting-state functional MR imaging duration on stability of graph theory metrics of brain network connectivity. Radiology 259:516–524. Wright IC, Ellison ZR, Sharma T, Friston KJ, Murray RM, McGuire PK (1999) Mapping of grey matter changes in schizophrenia. Schizophr Res 35:1–14. Wright IC, McGuire PK, Poline JB, Travere JM, Murray RM, Frith CD, Frackowiak RS, Friston KJ (1995) A voxel-based method for the

statistical analysis of gray and white matter density applied to schizophrenia. Neuroimage 2:244–252. Wysokinski A, Zboralski K, Orzechowska A, Galecki P, Florkowski A, Talarowska M (2010) Normalization of the Verbal Fluency Test on the basis of results for healthy subjects, patients with schizophrenia, patients with organic lesions of the chronic nervous system and patients with type 1 and 2 diabetes. Arch Med Sci 6:438–446.

(Accepted 11 November 2014) (Available online 26 November 2014)