Differences in network centrality between high and low myopia: a voxel-level degree centrality study

Abstract

Background

Previous studies have linked high myopia (HM) to brain activity, and the difference between HM and low myopia (LM) can be assessed.

Purpose

To study the differences in functional networks of brain activity between HM and LM by the voxel-level degree centrality (DC) method.

Material and Methods

Twenty-eight patients with HM (10 men, 18 women), 18 patients with LM (4 men, 14 women), and 59 healthy controls (27 men, 32 women) were enrolled in this study. The voxel-level DC method was used to assess spontaneous brain activity. Correlation analysis was used to explore the change of average DC value in different brain regions, in order to analyze differences in brain activity between HM and LM.

Results

DC values of the right cerebellum anterior lobe/brainstem, right parahippocampal gyrus, and left caudate in HM patients were significantly higher than those in LM patients (P < 0.05). In contrast, DC values of the left medial frontal gyrus, right inferior frontal gyrus, left middle frontal gyrus, and left inferior parietal lobule were significantly lower in patients with HM (P < 0.05). However, there was no correlation between behavior and average DC values in different brain regions (P < 0.05).

Conclusion

Different changes in brain regions between HM and LM may indicate differences in neural mechanisms between HM and LM. DC values could be useful as biomarkers for differences in brain activity between patients with HM and LM. This study provides a new method to assess differences in functional networks of brain activity between patients with HM and LM.

Keywords

High myopia low myopia degree centrality brain regions

Introduction

With the increasing use of electronic devices, myopia has become a primary public health problem worldwide. There are two types of myopia: high (HM) and low (LM). HM refers to myopia of ≥–6 diopter (D), whereas LM refers to myopia of ≤–3 D. HM is the more harmful type and its incidence is also increasing each year (1). HM generally constitutes axial myopia, where the axial length of the eye is prolonged such that proptosis and deep anterior chamber are present. Thus far, most treatment methods constitute symptomatic treatment, such as eyeglass wear, corneal surgery, and lens surgery; these change the diopter of the image that reaches the retina, thereby achieving the goal of correction. However, the excessive lengthening of the axial length can cause a series of pathological fundus changes, such as leopard-like fundus, choroidal atrophy around the optic disc, Foster-Fuchs spots, and posterior scleral staphyloma. Pathologic changes in the eyeball are reduced in LM, compared with HM, such that the harm is reduced. Optical coherence tomography (OCT) is a common imaging method used in ophthalmology, which can be used to detect the ocular fundus in patients with HM. In these patients, the retinal nerve fiber layer becomes thinner and the optic papilla becomes deformed (2). In a prospective study of HM using OCT, Lee et al. (3) found that the thickness of the retinal nerve fiber layer near the papilla in eyes with HM was significantly lower than that in normal eyes; moreover, the rate of reduction was higher in elderly patients with HM. Milani et al. (4) reported that superficial vascular density decreased in HM and was positively correlated with retinal thickness; therefore, the retinal thickness also decreased in HM. Retinal nerve fibers converge out of the ethmoidal layer of the sclera and enter the skull through the orbit, subsequently connecting with the brain. Because HM can lead to changes in the retinal nerve fiber layer, the brain areas connected to it may also be altered. Recently, an increasing number of magnetic resonance imaging (MRI) studies have been performed regarding the eye morphology in myopic patients (5); the results of those studies suggest that MRI findings may be associated with visual impairment (6). Research by Wen et al. (7) has shown that high-resolution three-dimensional (3D) MRI can be used to quantitatively analyze the morphology of eyes with HM. However, good visual function requires a complete eye shape, as well as a complete visual pathway and a normal visual cortex; thus, poor visual quality in patients with HM may be related to changes in the eye, as well as changes in the visual pathway and visual cortex. However, most MRI studies of HM are limited to eye morphology, and there has been minimal research on vision-related brain activity.

The application of functional MRI (fMRI) has considerably aided in understanding the functional and structural properties of the human visual cortex (8). Elbel et al. (9) proposed that myopia may change visual cortical activity. Mirzajani et al. (10) evaluated the effects of lens-induced HM on activity in the occipital visual cortex during the presentation of two visual stimuli fMRI responses to visual stimuli with a spatial frequency of 1.84 CPD; they reported a significant 10-fold reduction in visual cortical activity in lens-induced HM, compared to normal vision. Thus far, most studies on myopia focus on regional brain structure, such as the visual cortex, but not on whole-brain aspects. However, the human brain is a complex and well-organized functional network that requires the coordination of multiple brain areas to perform specific behaviors. Therefore, it is necessary to study the connectivity of the whole-brain network. In recent years, the development of graph theory has enabled improved understanding of functional brain networks, and graph theory analysis has become increasingly useful for the exploration of whole-brain functional organization and connectivity (11,12). Voxel level degree centrality (DC) is a graph-based and data-driven approach that assesses the importance of each voxel in the brain, such that voxels represent their connection strength (13). The DC index is a better measure of connectivity than other measurements, in this method, DC refers to the number of direct connections between a given voxel and the rest of the whole voxel, rather than the number of direct connections between specific nodes or regions. Voxel-wise DC allows us to map brain functional centers without prior nodes or regions of interest (brain regions are concentrated by a large number of connections with the rest of the brain). Centrality emphasizes the influence and importance of the network at the voxel level, which reflects the characteristics of the central region of the functional brain network in network information communication (14,15). In functional brain networks, the central region plays a key role in the coordination of information flow, which is synchronized and stable in the brain; however, the central region is highly susceptible to disease, which may result in a series of changes. Degree-centered measurements based on resting state fMRI have been used to observe changes in functional networks in different diseases, and these measurements have shown relatively high test–retest reliability. DC that focused on voxels may provide a breakthrough in the pathogenesis of myopia. In recent years, myopia has become an important topic in health research. We have performed in-depth analyses of HM and found that HM is associated with abnormalities in many brain regions, compared with normal eyes. Altered DC values can be used as biomarkers for brain activity changes in patients with HM (16). In addition, differences between HM and LM can be assessed. Notably, we found that there is minimal research comparing HM and LM in terms of DC values; thus, we performed this study to explore the neural mechanisms of HM and LM by studying whole-brain activity in patients with HM and in patients with LM, in order to identify novel targets for prevention and treatment of myopia.

Material and Methods

Participants

Participants were from a hospital during the period from September 2015 to September 2017. The cohort included 28 patients with HM, 18 patients with LM, and 59 healthy controls (HCs) without myopia. Patients were similar in sex, age, and level of education. The inclusion criteria for myopia patients were as follows: (i) aged 18–60 years; (ii) right-handed; (iii) presence of HM or LM; and (iv) absence of other eye diseases. Exclusion criteria were as follows: (i) presence of other ophthalmic diseases; (ii) presence of unilateral myopia; (iii) presence of abnormal brain structure, such as tumor or subdural hematoma; and (iv) presence of mental illness. The HC group was matched to patients with HM and patients with LM for age, sex, handedness, and level of education. If HCs met the following criteria, they were included in the study: (i) absence of refractive error and other eye diseases; (ii) absence of mental illness; and (iii) presence of normal brain MRI findings. In addition, all individuals underwent an MRI scan; they were excluded if they had incomplete MRI data or exhibited excessive head movement. The study was approved by the medical ethics committee and all participants provided informed consent.

MRI collection

The heads of individuals in the HM, LM, and HC groups were scanned by a Siemens 3-T scanner. Gradient echo images (MPRAGE) were prepared by the magnetization method with the following parameters: repetition time (TR) = 1900 ms; echo time (TE) = 2.26 ms; thickness = 1.0 mm; no intersection; acquisition matrix = 256 × 256; field of view (FOV) = 240 × 240 mm; and flip angle (FA) = 12°. Whole-brain T1-weighted (T1W) images were obtained. Functional images were obtained by a gradient echo plane imaging sequence with the following parameters: TR = 2000 ms; TE = 40 ms; FA = 90°; slice thickness/gap = 4.0/1 mm; FOV = 240 × 240 mm; and plane resolution = 64 × 64. Thirty axial sections and 240 volumes were obtained in 8 min.

Resting-state fMRI data preprocessing

Preprocessing was performed by using a toolbox for Data Processing & Analysis of Brain Imaging (DPABI, http://rfmri.org/dpabi) based on statistical parametric mapping (SPM12) running on MATLAB 8.4.0 (Mathworks, Natick, MA, USA). The main preprocessing steps were as follows: the first 10 volumes were discarded; slice timing, spatial realignment, head motion correction, and individual registration between high-resolution T1 and EPI images were performed. Then, spatial normalization was implemented to register resting-state fMRI datasets to the Montreal Neurological Institute (MNI) space; the data were re-sampled as 3 × 3 × 3 mm cube voxels; head motion estimation was performed and images with > 2.0 mm of maximal translation or >2.0° of maximal rotation were excluded from the final analysis. Preprocessing data were divided into typical frequency bands (0.01–0.1 Hz) for the calculation of DC. First, linear detrending and nuisance linear regression were performed; then, white matter signal, cerebrospinal fluid signal, global signal, and Friston 24 parameters are regressed from the time series of all voxels (40); third, typical frequency band: human brain is a complex dynamic system, which can generate low-frequency oscillation (LFO) at rest. LFO (0.01–0.08 Hz) on blood oxygen level-dependent (BOLD) imaging has been shown to play an important role in various physiological activities (41,42).

DC analysis

Individual DC mapping calculations were performed using the DC analysis module in DPABI. DC refers to the number of connections between a node or brain region and other nodes or brain regions in the whole-brain network; it reflects the attributes of complex brain network functional connections from a new perspective and can be used to observe changes in a patient’s brain network. The post-processing analysis method for DC in this study followed the methods of Buckner et al. (17) and Zuo et al. (13); key parameters were selected as follows: preprocessed fMRI data were not smoothed; and whole-brain information was retained when covariates were removed. When DC values were calculated, r was 0.25; the z-value weighted graph was used for statistical analysis. First, the voxel-wise DC was calculated by the using the DPABI software in the default gray matter mask provided by DPABI. Each voxel was taken as a node and inter-voxel correlations as the edge. The voxel-wise DC map for each individual was calculated using the following equation:

DC (i) = \sum_{j =1}^{N} r_{i j} (r_{i j} > r_{0})

where r_ij is the correlation coefficient between voxel i and voxel j, and r₀ is a correlation threshold that is set to eliminate the weak correlation. According to previous studies, we analyzed the different DC values with correlation threshold at (r₀ = 0.25) in the present study (13,17). Second, two groups of DC ROI passed the two-sample t-test. The two-sample t-test was used to compare the differences between the two groups of zDC maps using the Gaussian random field (GRF) method. It was used to correct for multiple comparisons, and spm8 software was used to regression the covariates and FD of age and gender (two-tailed, voxel-level P < 0.01, GRF correction, cluster-level P < 0.05). Third, node uses all brain area template in DPABI. Fourth, this regression covariate operation has been done in the previous preprocessing. Finally, individual DC maps underwent Fisher’s r-to-z standardization within whole brain mask; the resulting data were further spatially smoothed using a 6-mm isotropic full width at half-maximum Gaussian kernel.

Statistical analysis

One-sample t-test was conducted to investigate the group mean z-values of DC in patients HM or LM and HCs. To explore differences in typical DC frequency bands between patients and HCs, an ANOVA and second-level random-effects two-sample t-test were conducted to investigate different DC values across three group. (two-tailed P < 0.01, GRF theory correction with cluster-level P < 0.05). All significant clusters were reported on MNI coordinates and T-values of the peak voxel were determined. Clinical characteristics and index statistics were evaluated using two-sample t-tests and Chi-squared tests in SPSS (version 13.0, SPSS Inc., Chicago, IL, USA), to examine whether clinical indices (e.g. JOA and VAS) of the patients were related to the alteration of DC.

Results

Basic information

The basic examination results of participants are shown in Table 1.

Table 1.

General characteristics of patients.

	Age (years)	ALM (OD)	ALM (OS)	Average macular thickness (µm) (OD)	Average macular thickness (µm) (OS)
HM	25.22 ± 6.467	26.58 ± 0.985	26.50 ± 1.013	258.59 ± 21.925	266.46 ± 17.174
LM	24.78 ± 3.302	24.90 ± 0.771	24.74 ± 0.971	269.26 ± 20.260	268.09 ± 20.518

ALM, axial length; DC, degree centrality; HM, high myopia; LM, low myopia; OD, Oculus dexter; OS, Oculus sinister.

Spatial distribution patterns of DC in patients with myopia and healthy controls

At the group mean level, patients exhibited similar spatial distributions of DC, as shown in Fig. 1.

Fig. 1.

Distribution patterns of DC were observed at group levels in patients with HM (a), patients with LM (b), and HCs (c).

Disease-related differences in DC

ANOVA analysis of DC among the HM, LM, and HC groups is shown in Fig. 2 (P < 0.01, GRF correction) and Table 2. A post hoc two-sample t-test was used to explore differences in group comparisons (Fig. 3, P < 0.01, GRF correction, Fig. 4, P < 0.01, without correction, Tables 3 and 4). Compared with the HC group, the HM group showed increased DC in the bilateral superior frontal gyrus/supplementary motor area (SMA). In contrast, HM patients had significantly reduced DC values in the right parahippocampal/inferior frontal gyrus (IFG) and bilateral superior frontal gyrus. Compared with the HC group, LM patients showed increased DC in the left middle frontal gyrus (MFG) and decreased DC in the right parahippocampal gyrus. Most importantly, compared with patients with LM, patients with HM showed increased DC in the right cerebellum anterior lobe/brainstem, right parahippocampal gyrus, and left caudate; patients with HM also showed significantly reduced DC values in the left medial frontal gyrus (LMFG), right inferior frontal gyrus (RIFG), left MFG, and left inferior parietal lobule.

Fig. 2.

ANOVA of DC among HM, LM, and HC groups (voxel-level P < 0.01, GRF correction, cluster-level P < 0.05).

Table 2.

ANOVA of DC among HM, LM, and HC groups (voxel-level P < 0.01, GRF correction, cluster-level P < 0.05).

Brain regions	BA	Peak scores	MNI coordinates			Cluster size (voxels)
Brain regions	BA	Peak scores	x	y	z	Cluster size (voxels)
Right fusiform/parahippocampal gyrus	20,28	10.08	24	–9	–36	105
Left inferior temporal gyrus	20	10.98	–54	–30	–18	115
Left cuneus	17,18	11.87	–6	–99	0	81
Bilateral MCC/ACC	24,32	9.87	9	18	42	95
Bilateral superior frontal gyrus/paracentral lobule	5,6,7	10.61	–6	–12	75	238

ACC, Anterior Cingulate Cortex; BA, Brodmann area; DC, degree centrality; GRF, Gaussian random field; HC, healthy control; HM, high myopia; LM, low myopia; MCC, middle cingulate cortex; MNI, Montreal Neurological Institute.

Fig. 3.

Group comparisons of DC between HM and HC groups, and between LM and HC groups (voxel-level P < 0.01, GRF correction, cluster-level P < 0.05).

Fig. 4.

Significant disease-related differences in DC between HM and LM groups (voxel-level P < 0.01, without correction).

Table 3.

Significant disease-related differences in DC between HM and HC groups, and between LM and HC groups (voxel-level P < 0.01, GRF correction, cluster-level P < 0.05).

Brain regions	BA	Peak scores	MNI coordinates			Cluster size (voxels)
Brain regions	BA	Peak scores	x	y	z	Cluster size (voxels)
Altered DC (HM vs. HC)
Right parahippocampal/inferior frontal gyrus	47,28,20	−3.845	27	18	−21	259
Bilateral superior frontal gyrus	10,9	−4.441	6	69	9	307
Bilateral superior frontal gyrus/SMA	6	4.209	6	−42	81	241
Altered DC (LM vs. HC)
Right parahippocampal gyrus	35,28	−4.811	−27	−24	−24	335
Left middle frontal gyrus	6,5	4.407	−24	6	51	268

BA, Brodmann area; DC, degree centrality; GRF, Gaussian random field; HC, healthy control; HM, high myopia; LM, low myopia; MNI, Montreal Neurological Institute; SMA, supplementary motor area.

Table 4.

Significant disease-related differences in DC between HM and LM groups (voxel-level P < 0.01, without correction).

Brain regions	BA	Peak scores	MNI coordinates			Cluster size (voxels)
Brain regions	BA	Peak scores	x	y	z	Cluster size (voxels)
Altered DC (HM vs. LM)
Right cerebellum anterior lobe/brainstem		3.37	12	–36	–30	62
Right parahippocampal gyrus	28	3.39	15	–18	–15	62
Left medial frontal gyrus	11	–4.51	–6	66	–9	114
Left caudate		3.52	–6	3	–3	47
Right inferior frontal gyrus	46	–4.48	45	54	6	95
Left middle frontal gyrus	10,46	–3.69	–48	36	30	63
Left inferior parietal lobule	40	–4.57	–63	–36	33	49

BA, Brodmann area; DC, degree centrality; HM, high myopia; LM, low myopia; MNI, Montreal Neurological Institute.

Discussion

In the present study, we evaluated DC differences among groups, especially between patients with HM and patients with LM. We found that, compared with HCs, the DC value of the bilateral superior frontal gyrus/SMA increased in patients with HM. In contrast, the DC values of the right parahippocampal/IGF and bilateral superior frontal gyrus were significantly lower in patients with HM. Compared with HCs, the DC value increased in the left MFG of patients with LM, whereas the DC value of the right parahippocampal gyrus decreased. Most importantly, patients with HM showed increased DC in the right cerebellum anterior lobe/brainstem, right parahippocampal gyrus, and left caudate, compared to patients with LM. In contrast, patients with HM had significantly lower DC values in the LMFG, RIFG, left MFG, and left inferior parietal lobule.

The cerebellum has long been regarded as the source of body balance and exercise control. In addition, the cerebellum participates in language processing, advanced cognition, memory, attention, emotion, and other information components needed for automation of learning skills. Furthermore, the cerebellum has been shown to participate in accurate eye movement (18,19). In the present study, we found that the DC value of the right cerebellum anterior lobe/brainstem increased in patients with HM compared to patients with LM, which indicated that the influence of the HM and LM on this brain area in the process of disease development was different. Therefore, we speculate that HM may have a greater impact on the function of the right cerebellum anterior lobe/brainstem than LM.

The parahippocampal gyrus has an important relationship with cognition and emotion, and structural damage to this area can cause abnormalities in emotional and cognitive behavior. The parahippocampal region is an important component of perceptual vision; it relies on two different networks that divide ppa, and participates in spatial processing and memory (20). Chen et al. (21) found that the hippocampus and parahippocampal gyrus play important roles in visual function. Through comparing the data of HM and LM, we found that the DC value of the right parahippocampal gyrus of HM was higher than the LM, which indicated that there were some differences in the effects of HM and LM on the brain area. It may be related to the obvious abnormal visual experience and changes in eye structure in the patients with HM. It also proves that the right parahippocampal gyrus has a close relationship with vision.

In the brain, the basal ganglia/caudate nucleus is primarily responsible for human habitual action; moreover, the caudate nucleus is an important part of the brain’s learning and memory systems. The caudate nucleus acts to adjust the movement of the body and has an influence on sensory conduction. In the early stages of visual motor adaptation, the basal ganglia pathway may play a variety of potential roles, including promoting motor and/or cognitive adaptation (22). Seidler et al. (23) speculated that the involvement of the left caudate and the globus pallidus, combined with SMA, contributed to the adaptation of the action selection process (23). Increased DC values in the left caudate of patients with HM suggest that HM has a greater effect on the neural activity in this brain region. Although the caudate nucleus is associated with learning and memory, whether HM can affect the left caudate nucleus to change the patients’ behavior and memory function still needs further study.

The LMFG is located in the frontal lobe, and its fibrous connections include the visual, auditory, and somatosensory cortices. LMFG dysfunction is associated with multiple conditions, including depression, Parkinson’s disease, brain tumors, glaucoma, and other diseases. In our study, DC in the LMFG of patients with HM was less than that of patients with LM, which indicates that the activity of neurons in this area is different between HM and LM. It might suggest that there was a certain difference in the specific neural mechanism of HM and LM affecting the LMFG, and then it might affect its function.

In addition, the frontal lobes involve most psychological functions, including memory, language, intelligence, and personality. In particular, the cortex around the lateral fissure of the brain is closely related to language function; this is known as the “language band.” Previous studies have shown that the right inferior frontal cortex is associated with response inhibition (24). In game-based inhibition training for elderly individuals, behavioral effects were observed in the inhibition-training group with a shorter stop signal response time; this was reflected by the cortical thickness of the right subfrontal triangle. Thus, structural growth is a response to this game-based inhibition training (25). In a language study, Qi et al. (26) found that successful foreign language acquisition in adult humans requires correct IFG participation during the initial learning period. In addition, the LIFG constitutes the neural IFG that processes abstract concepts and information required for those concepts (27). Tesink et al. (28) found an increase in right subfrontal triangle activation in the autism spectrum disorders (ASD) group in an ASD study; this may be related to higher task requirements in ASD participants when integrating speaker features and spoken sentence content. We observed a reduction in the DC value of patients with HM. HM may affect IFG function, and there are thus dysfunctions in language and behavioral control. The HM research by Huang et al. (29) is consistent with our results, supporting this hypothesis.

The MFG is one of the cerebral gyri; it is located in the frontal lobe of the cerebral cortex, between the superior and inferior frontal sulci. The posterior portion of this gyrus has similar function in directing movement of the center of the head and eyes. In a study of patients with strabismic amblyopia, the amplitude of low-frequency fluctuation (ALFF) value of the left MFG decreased in the strabismus with amblyopia (SA) group, suggesting that the frontal eye field (FEF) function was impaired; however, FEF could initiate eye movements and affect its latency or accuracy (30). The MFG is regarded as a second language field, and the LMFG has been shown to be related to language generation (31). Richards et al. (32) found that activity in the bilateral medial frontal gyrus was associated with reading when they studied children with dyslexia. A study of Alzheimer’s disease found that functional changes in the left MFG had a significant effect on cognitive function (33). Honey et al. (34) presumed that MFG activity was related to memory. Our results suggest that the reduction in DC value in the left MFG of patients with HM may indicate impaired language reading and memory cognitive function in these patients, compared to patients with LM.

The inferior parietal lobule is located below the parietal sulcus in the parietal lobe and participates in language learning. It is presumably connected to action and perception, such that it is an important part of the brain network of imitation and learning (35). Previous studies have shown that the left inferior parietal lobe is a key brain region for the attention control network and is involved in executive function (36). Hua et al. (37) assessed the internal connectivity disorder pattern of the whole-brain function network in patients with chronic use of codeine-containing cough syrups (CCS) at the voxel level; notably, they found that the degrees of reduction in the left inferior parietal lobe and left middle temporal gyrus were associated with cognitive function in individuals dependent on CCS. Wang et al. (38) explored the functional network of brain activity changes in patients with acute unilateral open globe injury and its relationship with clinical characteristics; they found that the DC values of right insula, left insula, right inferior parietal lobule (IPL)/supramarginal gyrus (SMG), left IPL/SMG, right SMA, and S1 were reduced, indicating that these brain changes were related to visual injury. In addition, in a study of speech perception, left IPL and left superior temporal gyrus (LSTG) were found to be associated with two key dimensions (pitch and pitch direction), whereas right STG was not. In addition, during tone classification, the degrees of IFG and parietal lobe activation were higher for similar stimuli, which reveals the role of left IPL in language perception (39). These studies also support our findings that the DC value of the left IPL was reduced in patients with HM compared to patients with LM, revealing a certain influence on language perception and cognitive function in patients with HM.

The present has some limitations. Because the participants in this study are young patients, the results may not represent the whole population, but selecting young patients can exclude the influence of some old brain changes of the elderly on the experiment, and also remove the children’s incompatibility and uncertainty. In addition, there may be some limitations in MRI, such as the head movement and cerebrospinal fluid movement effect during the patient’s examination. However, we strictly select the qualified examination results and remove the influence of the motion effect produced in the brain by statistical methods, in order to minimize the influence of these factors on the experimental results.

In conclusion, compared with patients with LM, the DC values of different brain regions in patients with HM have different changes, suggesting that HM and LM can affect the neural activities of the right cerebellum anterior lobe/brainstem, right parahippocampal gyrus, left caudate, the LMFG, RIFG, left middle frontal gyrus, and left inferior parietal lobule. Through the changes of DC values in different brain regions, we can find that HM and LM have different effects on these brain regions, which provides a new breakthrough for exploring the difference of neural mechanism between HM and LM, and may provide a new direction for the diagnosis and treatment of HM and LM.

Footnotes

Authors' note

Xiaorong Wu (wxr98021@126.com) and Yu Tian (tianyu3734358@csu.edu.cn) are corresponding authors.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received the following financial support for the research, authorship, and/or publication of this article: National Natural Science Foundation of China (No. 81760179, 81360151); Natural Science Foundation of Jiangxi Province (No. 20171BAB205046); Jiangxi Province Education Department Key Foundation (No. GJJ160033); Health Development Planning Commission Science Foundation of Jiangxi Province (No. 20185118); Technology and Science Foundation of Jiangxi Province (No. 20141BBG70027); Jiangxi Province Education Department Scientific Research Foundation (No. GJJ13147); Health and Family Planning Commission Traditional Chinese Medicine Foundation of Jiangxi Province (No. 20185118, 20141031); and Basic Health Appropriate Technology Spark Promotion Program of Jiangxi Province (No. 20188007).

ORCID iDs

Xin Huang

Xiaorong Wu

References

Cheung

Luk

, et al. Advances of optical coherence tomography in myopia and pathologic myopia. Eye (Lond) 2016; 30:901–916.

Lin

Jiang

, et al. Retinal nerve fiber layer (RNFL) integrity and its relations to retinal microvasculature and microcirculation in myopic eyes. Eye Vis (Lond) 2018; 5:25.

Lee

Kim

Shin

, et al. Longitudinal changes in peripapillary retinal nerve fiber layer thickness in high myopia: a prospective, observational study. Ophthalmology 2019; 126:522–528.

Milani

Montesano

Rossetti

, et al. Vessel density, retinal thickness, and choriocapillaris vascular flow in myopic eyes on OCT angiography. Graefes Arch Clin Exp Ophthalmol 2018; 256:1419–1427.

Liu

, et al. Morphological analysis and quantitative evaluation of myopic maculopathy by three-dimensional magnetic resonance imaging. Eye (Lond) 2018; 32:782–787.

Moriyama

Ohno-Matsui

Hayashi

, et al. Topographic analyses of shape of eyes with pathologic myopia by high-resolution three-dimensional magnetic resonance imaging. Ophthalmology 2011; 118:1626–1637.

Wen

Yang

Cheng

, et al. Using high-resolution 3D magnetic resonance imaging to quantitatively analyze the shape of eyeballs with high myopia and provide assistance for posterior scleral reinforcement. Ophthalmologica 2017; 238:154–162.

Kollias

SS.

Investigations of the human visual system using functional magnetic resonance imaging (FMRI). Eur J Radiol 2004; 49:64–75.

Elbel

Kaufmann

Schaefers

, et al. Refractive anomalies and visual activation in functional magnetic resonance imaging (fMRI): a versatile and low-cost MR-compatible device to correct a potential confound. J Magn Reson Imaging 2002; 15:101–107.

10.

Mirzajani

Ghorbani

Rasuli

, et al. Effect of induced high myopia on functional MRI signal changes. Phys Med 2017; 37:32–36.

11.

Strogatz

SH.

Exploring complex networks. Nature 2001; 410:268–276.

12.

Bullmore

Sporns

Complex brain networks: graph theoretical analysis of structural and functional systems. Nat Rev Neurosci 2009; 10:186–198.

13.

Zuo

Ehmke

Mennes

, et al. Network centrality in the human functional connectome. Cereb Cortex 2012; 22:1862–1875.

14.

Telesford

Simpson

Burdette

, et al. The brain as a complex system: using network science as a tool for understanding the brain. Brain Connect 2011; 1:295–308.

15.

Tomasi

Volkow

ND.

Functional connectivity hubs in the human brain. Neuroimage 2011; 57:908–917.

16.

Yang

, et al. Abnormal resting-state functional network centrality in patients with high myopia: evidence from a voxel-wise degree centrality analysis. Int J Ophthalmol 2018; 11:1814–1820.

17.

Buckner

Sepulcre

Talukdar

, et al. Cortical hubs revealed by intrinsic functional connectivity: mapping, assessment of stability, and relation to Alzheimer’s disease. J Neurosci 2009; 29:1860–1873.

18.

Buckner

RL.

The cerebellum and cognitive function: 25 years of insight from anatomy and neuroimaging. Neuron 2013; 80:807–815.

19.

Timmann

Daum

Cerebellar contributions to cognitive functions: a progress report after two decades of research. Cerebellum 2007; 6:159–162.

20.

Aminoff

Gronau

Bar

The parahippocampal cortex mediates spatial and nonspatial associations. Cereb Cortex 2007; 17:1493–1503.

21.

Chen

Zheng

Chen

, et al. Whether visual-related structural and functional changes occur in brain of patients with acute incomplete cervical cord injury: a multimodal based MRI study. Neuroscience 2018; 393:284–294.

22.

Middleton

Strick

PL.

Basal ganglia and cerebellar loops: motor and cognitive circuits. Brain Res Brain Res Rev 2000; 31:236–250.

23.

Seidler

Noll

Chintalapati

Bilateral basal ganglia activation associated with sensorimotor adaptation. Exp Brain Res 2006; 175:544–555.

24.

Cunillera

Brignani

Cucurell

, et al. The right inferior frontal cortex in response inhibition: A tDCS-ERP co-registration study. Neuroimage 2016; 140:66–75.

25.

Kuhn

Lorenz

Weichenberger

, et al. Taking control! Structural and behavioural plasticity in response to game-based inhibition training in older adults. Neuroimage 2017; 156:199–206.

26.

Han

Wang

, et al. Speech processing and plasticity in the right hemisphere predict variation in adult foreign language learning. Neuroimage 2019; 192:76–87.

27.

Della

Catricala

Canini

, et al. The left inferior frontal gyrus: A neural crossroads between abstract and concrete knowledge. Neuroimage 2018; 175:449–459.

28.

Tesink

Buitelaar

Petersson

, et al. Neural correlates of pragmatic language comprehension in autism spectrum disorders. Brain 2009; 132:1941–1952.

29.

Zhou

, et al. Altered whole-brain gray matter volume in high myopia patients: a voxel-based morphometry study. Neuroreport 2018; 29:760–767.

30.

Min

Shu

, et al. Altered spontaneous brain activity patterns in strabismus with amblyopia patients using amplitude of low-frequency fluctuation: a resting-state fMRI study. Neuropsychiatr Dis Treat 2018; 14:2351–2359.

31.

Gauvin

McMahon

Meinzer

, et al. The shape of things to come in speech production: a functional magnetic resonance imaging study of visual form interference during lexical access. J Cogn Neurosci 2019; 31:913–921.

32.

Richards

Berninger

Yagle

, et al. Brain’s functional network clustering coefficient changes in response to instruction (RTI) in students with and without reading disabilities: Multi-leveled reading brain’s RTI. Cogent Psychol 2018; 5:1424680.

33.

Sun

Chen

Zuo

, et al. Effect of PICALM rs3851179 polymorphism on the default mode network function in mild cognitive impairment. Behav Brain Res 2017; 331:225–232.

34.

Honey

Kim

, et al. Effects of verbal working memory load on corticocortical connectivity modeled by path analysis of functional magnetic resonance imaging data. Neuroimage 2002; 17:573–582.

35.

Rizzolatti

Ferrari

Rozzi

, et al. The inferior parietal lobule: where action becomes perception. Novartis Found Symp 2006; 270:129–140, 140–145, 164–169.

36.

Monge

Geib

Siciliano

, et al. Functional modular architecture underlying attentional control in aging. Neuroimage 2017; 155:257–270.

37.

Hua

Wang

, et al. Abnormal degree centrality in chronic users of codeine-containing cough syrups: A resting-state functional magnetic resonance imaging study. Neuroimage Clin 2018; 19:775–781.

38.

Wang

Chen

, et al. Network centrality in patients with acute unilateral open globe injury: A voxelwise degree centrality study. Mol Med Rep 2017; 16:8295–8300.

39.

Feng

Gan

Wang

, et al. Task-general and acoustic-invariant neural representation of speech categories in the human brain. Cereb Cortex 2018; 28:3241–3254.

40.

Friston

Williams

Howard

, et al. Movement-related effects in fMRI time-series. Magn Reson Med 1996; 35:346–355.

41.

Wise

Ide

Poulin

, et al. Resting fluctuations in arterial carbon dioxide induce significant low frequency variations in BOLD signal. Neuroimage 2004; 21:1652–1664.

42.

Chan

Leong

ATL

, et al. Low-frequency hippocampal-cortical activity drives brain-wide resting-state functional MRI connectivity. Proc Natl Acad Sci USA. 2017; 114:E6972–E6981