Differential patterns of interhemispheric functional connectivity between AQP4-optic neuritis and MOG-optic neuritis: a resting-state functional MRI study

Abstract

Background

Several neuroimaging studies demonstrated that optic neuritis (ON) leads to functional and anatomical architecture changes in the brain. The alterations of interhemispheric functional connectivity (IFC) in patients with AQP4-ON and myelin oligodendrocyte glycoprotein (MOG)-ON are not well understood.

Purpose

To investigate the differential patterns of VMHC in patients with AQP4-ON and MOG-ON.

Material and Methods

Twenty-one patients with AQP4-ON, 11 patients with MOG-ON, and 34 healthy controls underwent resting-state MRI scans. One-way ANOVA was used to identify regions in which the zVMHC differed among the three groups. Post hoc two-sample t-tests were then conducted to compare zVMHC values between pairs of groups. Pearson correlation analysis was conducted to reveal relationships between mean zVMHC values and clinical variables in the AQP4-ON and MOG-ON groups.

Results

The results revealed significant differences in zVMHC values in the PreCG among the three groups. Compared to the control group: the AQP4-ON group showed significantly lower VMHC values in the superior temporal gyrus, inferior frontal gyrus, and PreCG; and the MOG-ON group showed significantly higher zVMHC values in the PostCG. Compared to the AQP4-ON group, the MOG-ON group showed significantly lower zVMHC values in the PreCG/PostCG (voxel-level P<0.01, GRF correction, cluster-level P<0.05)

Conclusion

Patients with AQP4-ON and those with MOG-ON showed abnormal VMHC in the motor cortices, sensorimotor cortices, and frontal lobe, possibly indicating impaired sensorimotor function in patients with ON. Moreover, differential patterns of VMHC in patients with AQP4-ON, compared to patients with MOG-ON, might serve as a clinical indicator for classification of ON.

Keywords

Anti-aquaporin-4 antibody anti-myelin oligodendrocyte glycoprotein antibody optic neuritis resting-state functional magnetic resonance imaging voxel-mirrored homotopic connectivity interhemispheric functional connectivity

Introduction

Optic neuritis (ON) is an inflammatory demyelination disease of the optic nerves, characterized by acute visual loss and eye pain during movement. ON can occur in isolation or in combination with other diseases, such as multiple sclerosis (MS) or neuromyelitis optica (NMO) spectrum disorder (1 –3). Anti-aquaporin-4 (AQP4) antibody was the first validated diagnostic biomarker, with a diagnostic specificity of approximately 91% for NMO (4,5). There is evidence that some patients with NMO are seronegative for anti-AQP4-antibody, whereas they are seropositive for anti-myelin oligodendrocyte glycoprotein (MOG) antibody (6). In recent years, there has been increasing awareness of the potential value of anti-MOG antibody in differentiating between ON phenotypes (7,8). In previous neuroimaging studies, patients with ON exhibited brain magnetic resonance imaging (MRI) abnormalities (9 –11). Furthermore, patients with ON exhibited changes in brain activity. Ren et al. (12) demonstrated that patients with single and relapsing ON both exhibited significantly lower amplitude of low-frequency fluctuations in the occipital and temporal lobes, relative to healthy controls (HCs). Huang et al. (13) reported that patients with single attack ON exhibited lower functional connectivity values in the frontal lobe, temporal lobe, right inferior occipital gyrus, right insula, and right inferior parietal lobule, compared to HCs; these patients also exhibited significantly higher functional connectivity values in the left thalamus. Gallo et al. (14) reported that patients with relapsing-remitting MS showed lower functional connectivity in the peristriate visual cortex. In addition, compared to patients with MS who did not exhibit ON, patients with ON-MS exhibited higher functional connectivity in the extrastriate cortex, at the level of right lateral middle occipital gyrus; they also exhibited lower functional connectivity at the level of the right inferior peristriate cortex. The abovementioned studies demonstrated changes in local brain activity in patients with ON; however, it remains unknown whether interhemispheric connectivity changes in these patients.

The human brain comprises two symmetric cerebral hemispheres with extensive anatomical and functional connections between homotopic locations. The synchrony of homotopic connectivity is an important feature of the functional structure of resting-state functional MRI (rs-fMRI); based on blood oxygenation level-dependent signals, it has been successfully used in vivo to reveal patterns of interhemispheric connectivity in the brain (15). Voxel-mirrored homotopic connectivity (VMHC) is an established rs-fMRI approach that quantifies the homotopic connectivity between hemispheres (16). Notably, the VMHC method has demonstrated high test–retest reliability. The VMHC method has been successfully used to investigate abnormal homotopic connectivity in various diseases, such as primary open-angle glaucoma (1), amblyopia (17), and Alzheimer's disease (AD) (18). However, it remains unknown whether abnormal interhemispheric connectivity is present in patients with ON.

The aim of the present study was to investigate whether patients with AQP4-IgG-positive ON (AQP4-ON) and patients with MOG-IgG-positive ON (MOG-ON) both exhibited abnormal interhemispheric connectivity. Furthermore, the study examined whether differential patterns of interhemispheric functional connectivity occur in patients with AQP4-ON compared to patients with MOG-ON. The results may offer insight into understanding of abnormal interhemispheric communication in patients with ON.

Material and Methods

Participants

The study protocol was approved by the Medical Ethics Committee and Department of Ophthalmology. In total, 32 right-handed patients with ON (i.e. 21 with AQP4-ON and 11 with MOG-ON) were recruited from the Department of Ophthalmology. The inclusion criteria for patients with AQP-ON and MOG-ON were as follows: (i) extended visual evoked potential P100 latency periods (bilateral or unilateral); (ii) acute ON; (iii) optic nerve MRI with T2-weighted hyper-intense lesion or T1-weighted (T1W) gadolinium-enhancing lesion extending over half of the optic nerve length or involving the optic chiasma; (iv) seropositivity for AQP4-IgG or MOG-IgG. The exclusion criteria for patients with AQP4-ON and MOG-ON were as follows: (i) any evidence of compressive, ischemic, toxic, genetic, metabolic, or invasive optic neuropathy; (ii) acute vision loss due to retinal disease, sympathetic ophthalmia, or nervous system disease; (iii) obvious abnormality in brain parenchyma, as detected by brain MRI; (iv) congenital or acquired diseases (e.g. psychiatric disorder, hypertension, diabetes mellitus, or coronary artery disease) or addictions (e.g. heroin, smoking, or alcohol); (v) history of organ transplantation; (vi) extreme underweight or overweight (body mass index <18.5 kg/m² or > 24.9 kg/m²).

In total, 34 right-handed HCs (11 men, 23 women) were also recruited for the present study; all HCs were age- and status-matched to participants in the AQP4-ON and MOG-ON groups. HCs met the following criteria: (i) no ocular disease that resulted in uncorrected decimal visual acuity > 1.0; (ii) no psychiatric disorders (e.g. depression, bipolar disorder, or sleep disorders); and (iii) ability to undergo MRI scanning (e.g. no cardiac pacemaker or implanted metal devices).

The present study was performed in accordance with the principles of the Declaration of Helsinki. All participants were adults; they were informed of the purpose, methods, and potential risks entailed in the study, then provided their written informed consent.

MRI parameters

MRI scanning was performed on a 3-T MR scanner (Trio; GE Healthcare Europe GmbH, Freiburg, Germany). Whole-brain T1W images were obtained with a spoiled gradient-recalled echo sequence using the following parameters: TR = 1900 ms; TE = 2.26 ms; thickness = 1.0 mm; gap = 0.5 mm; acquisition matrix = 256 × 256; field of view (FOV) = 250 × 250 mm; and flip angle (FA) = 9°. Functional images were collected with the following parameters: TR = 2000 ms; TE = 30 ms; thickness = 4.0 mm; gap = 1.2 mm; acquisition matrix = 64 × 64: FA = 90°; FOV = 220 × 220 mm; and 29 axial slices.

fMRI data preprocessing

All preprocessing was performed using the Data Processing & Analysis of Brain Imaging (DPABI, http://www.rfmri.org/dpabi) toolbox (19), which is based on Statistical Parametric Mapping, version 8 (http://www.fil.ion.ucl.ac.uk) implemented in MATLAB 2013a (MathWorks, Natick, MA, USA). Briefly preprocessing was performed as follows: first, images in DICOM format were converted to NIFTI format, and the first 10 volumes of each participant were discarded to ensure signal had reached equilibrium in the analyzed images. Second, the remaining 230 volumes of functional blood oxygenation level-dependent images were corrected for slice timing effects, then motion-corrected and realigned. Data from participants with head motion >2 mm or for whom rotation exceeded 2° during scanning were excluded, in accordance with the criteria of Van Dijk et al. (20). Third, individual T1W MPRAGE structural images were registered to the mean fMRI data; resulting aligned T1W images were segmented using the Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra toolbox for improved spatial precision in the normalization of fMRI data (21). Normalized data (in Montreal Neurological Institute 152 space) were re-sliced at a resolution of 3 × 3 × 3 mm. Fourth, smoothing was performed with a 6-mm full-width-half-maximum Gaussian kernel. Fifth, linear regression analysis was used to regress out several covariates: six head motion parameters; mean framewise displacement; global brain signal; and averaged signal from white matter signal and cerebrospinal fluid. Finally, data with linear trend were removed, and temporal band-pass filtering (0.01–0.08 Hz) was performed. Global signal regression is a widely used preprocessing step in rsfMRI that has been shown to improve the spatial specificity of the estimated resting-state networks. Thus, global brain signal was regressed out in our study.

Voxel-mirrored homotopic connectivity analysis

VMHC was performed using the DPABI toolkit. A group-specific symmetrical left–right hemisphere template was generated from the normalized T1 image of all participants to minimize geometric differences between hemispheres. This transformation of the T1W image was then applied to each individual preprocessed resting-state functional image. For each participant, the homotopic connectivity coefficient was computed as the Pearson correlation between each pair of mirrored interhemispheric voxels’ time series. Subsequently, correlation values were Fisher z-transformed to improve normality.

Statistical analysis

One-sample t-tests were conducted to assess intra-group patterns of zVMHC maps using Statistical Parametric Mapping, version 8. One-way analysis of variance (ANOVA) was used to identify regions in which the zVMHC differed among the three groups using the DPABI toolkit (two-tailed, voxel-wise P < 0.01; Gaussian random field corrected, cluster level P < 0.05). Post hoc two-sample t-tests were then conducted to compare zVMHC values between pairs of groups using the DPABI toolkit (two-tailed, voxel-wise P < 0.01, Gaussian random field corrected, cluster level P < 0.05); age, sex, and framewise displacement were treated as covariates.

Pearson correlation analysis was conducted to reveal relationships between zVMHC values and clinical variables in patients with AQP4-ON and patients with MOG-ON using SPSS Statistics, version 20.0 (IBM Corp., Armonk, NY, USA).

Results

Demographics characteristics

Participant details are shown in Table 1, including 21 patients with AQP4-ON (21 women), 11 patients with MOG-ON (4 men, 7 women) and 34 HCs (11 men, 23 women). There were no significant differences among the three groups in terms of demographic characteristics.

Table 1.

Demographics characteristics across three groups.

	AQP4-ON group (n = 21)	MOG-ON group (n = 11)	HC group (n = 34)	P values
Gender (M/F)*	0/21	4/7	11/23	0.010
Age (years)^†	37.00 ± 8.79	32.18 ± 3.99	36.33 ± 6.23	0.154
Right-handed	21	11	34	N/A

Values are given as n or mean ± SD.

*χ² test.

^†Independent t-test.

AQP4, aquaporin-4; HC, healthy control; MOG, myelin oligodendrocyte glycoprotein; ON, optic neuritis.

VMHC group differences

Spatial patterns of VMHC at the group mean level in the AQP4-ON, MOG-ON, and HC groups are shown in Fig. 1. One-way ANOVA revealed significant differences in zVMHC values in the precentral gyrus (PreCG) among the three groups (Fig. 2 and Table 2). Compared to the HC group, the AQP4-ON group showed significantly lower zVMHC values in the superior temporal gyrus (STG), inferior frontal gyrus (IFG), and PreCG (Fig. 3 and Table 2). Compared to the AQP4-ON group, the MOG-ON group showed significantly lower zVMHC values in the PreCG/postcentral gyrus (PostCG) (Fig. 4 and Table 2). In addition, the MOG-ON group showed significantly higher zVMHC values in the PostCG, relative to the HC group (Fig. 5 and Table 2).

Fig. 1.

Results of a one-sample t-test. Note: Within-group VMHC maps within the AQP4-ON (left) and HCs (middle) and MOG-ONs (right) AQP4, aquaporin-4; HC, healthy control; MOG, myelin oligodendrocyte glycoprotein; ON, optic neuritis; VMHC, voxel-mirrored homotopic connectivity.

Fig. 2.

Results of a one-way ANOVA across three groups. Note: A one-way ANOVA was used to identify zVMHC difference across three groups. (two-tailed, voxel-wise P < 0.01, GRF theory connected, cluster-level, P < 0.05). ANOVA, analysis of variance; AQP4, aquaporin-4; GRF, Gaussian random field; HC, healthy control; MOG, myelin oligodendrocyte glycoprotein; ON, optic neuritis; PreCG, precentral gyrus; VMHC, voxel-mirrored homotopic connectivity.

Table 2.

Brain areas with significantly different VMHC values across three groups.

Brain regions	BA	Peak T-scores*	MNI coordinates (x, y, z)^†	Cluster size (voxels)
Results of one-way ANOVA across three groups
PreCG	3	5.3261	±63, –21, 6	2010
AQP4-ON vs. HC
STG	–	–4.7841	±63, –21, 6	663
IFG		–3.7482	±45, 24, 6	121
PreCG	3	–4.2948	±36, –9, 51	450
AQP4-ON vs. MOG-ON
PreCG/PostCG	3,4	–5.6145	±63, –18, 18	3541
MOG-ON vs. HC
PostCG	4	4.2725	±45, –27, −3	475

*Statistical value of peak voxels showing significant differences between the two groups.

^†x, y, z are the coordinates of primary peak locations in the MNI space (two-tailed, voxel-level P < 0.01, GRF correction, cluster-level P < 0.05).

ANOVA, analysis of variance; AQP4, aquaporin-4; BA, Brodmann’s area; GRF, Gaussian random field; HC, healthy control; IFG, inferior frontal gyrus; MNI, Montreal Neurological Institute; MOG, myelin oligodendrocyte glycoprotein; ON, optic neuritis; PostCG, postcentral gyrus; PreCG, precentral gyrus; STG, superior temporal gyrus; VMHC, voxel-mirrored homotopic connectivity.

Fig. 3.

Results of a post hoc two-sample t-test between AQP4-ON group and HC group. Note: A post hoc two-sample t-test was used to identify zVMHC between AQP4-ON group and HC group (two-tailed; voxel-wise P < 0.01, GRF theory connected, cluster-level, P < 0.05). AQP4, aquaporin-4; GRF, Gaussian random field; HC, healthy control; IFG, inferior frontal gyrus; ON, optic neuritis; PreCG, precentral gyrus; STG, superior temporal gyrus; VMHC, voxel-mirrored homotopic connectivity.

Fig. 4.

Results of a post hoc two-sample t-test between AQP4-ON group and MOG-ON group. Note: A post hoc two-sample t-test was used to identify zVMHC between AQP4-ON group and MOG-ON group (two-tailed; voxel-wise P < 0.01, GRF theory connected, cluster-level, P < 0.05). AQP4, aquaporin-4; GRF, Gaussian random field; HC, healthy control; MOG, myelin oligodendrocyte glycoprotein; ON, optic neuritis; PreCG, precentral gyrus; VMHC, voxel-mirrored homotopic connectivity.

Fig. 5.

Results of a post hoc two-sample t-test between MOG-ON group and HC group. Note: A post hoc two-sample t-test was used to identify zVMHC between MOG-ON group and HC group (two-tailed; voxel-wise P < 0.01, GRF theory connected, cluster-level, P < 0.05). GRF, Gaussian random field; HC, healthy control; MOG, myelin oligodendrocyte glycoprotein; ON, optic neuritis; PreCG, precentral gyrus; VMHC, voxel-mirrored homotopic connectivity.

Discussion

The VMHC method is a reliable and non-invasive fMRI technique that can be used to assess the homotopic connectivity architecture of the brain. To the best of our knowledge, the present study is the first to assess alterations in homotopic connectivity in patients with AQP4-ON and patients with MOG-ON, using the VMHC method. Notably, the AQP4-ON group showed significantly lower zVMHC values in the STG, IFG, and PreCG, compared to the HC group. Furthermore, the MOG-ON group showed significantly lower zVMHC values in the PreCG/PostCG, compared to the AQP4-ON group. Finally, the MOG-ON group showed significantly higher zVMHC values in the PostCG, relative to the HC group.

The PreCG, also known as the primary motor cortex, plays an important role in motor control, as well as in pain sensory processing. PreCG dysfunction has been observed in patients with complex regional pain syndrome (22). A neuroimaging study revealed that acute muscle orofacial and cutaneous pain were strongly associated with signal intensity reductions within the contralateral PreCG (23). The present study showed significantly different zVMHC values in the PreCG among the three groups; the AQP4-ON group exhibited lower VMHC values in the PreCG, while the MOG-ON group exhibited higher VMHC values in the PreCG. These findings may explain the occurrence of pain in patients with ON (e.g. eye pain, headache, or orbital pain). We hypothesized that differences in the VMHC of distinct brain regions among the AQP4-ON, MOG-ON, and HC groups could be useful as diagnostic markers. In addition, AQP4 may be associated with the incidence of amyotrophic lateral sclerosis, a progressive neurodegenerative disorder that targets motor neurons, causing muscle weakness and eventual paralysis (24). Another study found that eye movement abnormalities were common in AQP4-IgG-positive patients with NMO spectrum disorder (12). These findings were consistent with our results, which suggest that AQP4 might be involved in motor dysfunction during the onset of neurological disease.

In classical neurology, the posterior STG is regarded as the site of word recognition (25). The left IFG (pars orbitalis) was also implicated in several early imaging studies of semantic processing; a meta-analysis revealed critical involvement of the anteroventral left IFG in semantic processing (26). Compared to the HC group, the AQP4-ON group showed significantly lower zVMHC values in the STG and IFG. Thus, the inter-hemispheric homotopic connectivities of the STG and IFG may be suitable neuroimaging markers of AQP4-ON.

The primary somatosensory cortex is located in the PostCG, which is involved in encoding touch and pain (27,28); the PostCG is also involved in modulation of vision (29,30). Qin et al. (31) demonstrated that patients with blindness exhibited lower long-range functional connectivity density in primary somatosensory cortices, compared to HCs. In the present study, the MOG-ON group showed significantly higher zVMHC values in the PostCG, relative to the HC group; furthermore, the MOG-ON group showed significantly lower zVMHC values in the PreCG/PostCG, compared to the AQP4-ON group. A previous immunohistochemistry study demonstrated that AQP4 is expressed in sensory ganglia, such as trigeminal ganglia and dorsal root ganglia in the peripheral nervous system; AQP4 is exclusively localized to satellite glial cells surrounding the cell bodies of primary afferent sensory neurons in sensory ganglia (32). However, the pathophysiological relevance of AQP4 in somatosensory perception remains unclear. The findings of the present study may provide new insight regarding the involvement of water homeostasis in the peripheral sensory system. We presume that this explains why patients with MOG-ON have a better prognosis than patients with AQP4-ON.

The present study has some limitations. First, VMHC methods were used to investigate interhemispheric connectivity changes in patients with AQP4-ON and patients with MOG-ON. In future studies, multimodal neuroimaging methods should be used to illustrate alterations in brain activity. Second, the sample size was relatively small, which limits the generalizability of the findings and may have introduced a degree of bias in the results.

In conclusion, the results of the present study demonstrated that patients with AQP4-ON and patients with MOG-ON showed abnormal interhemispheric functional connectivity in the motor cortices, sensorimotor cortices, and frontal lobe, which might indicate impaired sensorimotor function in patients with ON. Moreover, differential patterns of interhemispheric functional connectivity observed in patients with AQP4-ON, compared to patients with MOG-ON, might serve as a clinical indicator for classification of ON.

Footnotes

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: The study was funded by the National Natural Science Foundation of China project “The mechanism research that CXCL12 promoted central nervous system myelination regeneration by activating JAK/STAT signal pathway” (no. 81870662).

ORCID iD

Mingming Sun

References

Wang

Chen

Wang

, et al. Reduced Functional and Anatomic Interhemispheric Homotopic Connectivity in Primary Open-Angle Glaucoma: a Combined Resting State-fMRI and DTI Study. Invest Ophthalmol Vis Sci 2018; 59:1861–1868.

Petzold

Plant

GT.

Diagnosis and classification of autoimmune optic neuropathy. Autoimmun Rev 2014; 13:539–545.

Toosy

Mason

Miller

DH.

Optic neuritis. Lancet Neurol 2014; 13:83–99.

Lennon

Wingerchuk

Kryzer

, et al. A serum autoantibody marker of neuromyelitis optica: distinction from multiple sclerosis. Lancet 2004; 364:2106–2112.

Hayreh

SS.

Ischemic optic neuropathies - where are we now?

Graef Arch Clin Exp 2013; 251:1873–1884.

Kitley

Waters

Woodhall

, et al. Neuromyelitis optica spectrum disorders with aquaporin-4 and myelin-oligodendrocyte glycoprotein antibodies: a comparative study. JAMA Neurol 2014; 71:276–283.

Kupersmith

Garvin

Wang

, et al. Retinal Ganglion Cell Layer Thinning Within One Month of Presentation for Non-Arteritic Anterior Ischemic Optic Neuropathy. Invest Ophth Vis Sci 2016; 57:3588–3593.

Kezuka

Ishikawa

Diagnosis and treatment of anti-myelin oligodendrocyte glycoprotein antibody positive optic neuritis. Jpn J Ophthalmol 2018; 62:101–108.

Heussinger

Kontopantelis

Rompel

, et al. Predicting multiple sclerosis following isolated optic neuritis in children. Eur J of Neurol 2013; 20:1292–1296.

10.

Hauser

Oksenberg

Lincoln

, et al. Interaction between HLA-DR2 and abnormal brain MRI in optic neuritis and early MS. Am J Ophthalmol 2000; 130:690–691.

11.

Orman

Wang

Pekcevik

, et al. Enhancing Brain Lesions during Acute Optic Neuritis and/or Longitudinally Extensive Transverse Myelitis May Portend a Higher Relapse Rate in Neuromyelitis Optica Spectrum Disorders. AJNR Am J Neuroradiol 2017; 38:949–953.

12.

Sun

Cui

Gao

, et al. Eye movement abnormalities in AQP4-IgG positive neuromyelitis optica spectrum disorder. J Neurol Sci 2018; 384:91–95.

13.

Huang

Duan

Liu

, et al. Altered Brain Structure and Functional Connectivity of Primary Visual Cortex in Optic Neuritis. Front Hum Neurosci 2018; 12:473.

14.

Gallo

Esposito

Sacco

, et al. Visual resting-state network in relapsing-remitting MS with and without previous optic neuritis. Neurology 2012; 79:1458–1465.

15.

Stark

Margulies

Shehzad

, et al. Regional variation in interhemispheric coordination of intrinsic hemodynamic fluctuations. J Neurosci 2008; 28:13754–13764.

16.

Zuo

Kelly

Di Martino

, et al. Growing together and growing apart: regional and sex differences in the lifespan developmental trajectories of functional homotopy. J Neurosci 2010; 30:15034–15043.

17.

Liang

Xie

Yang

, et al. Altered interhemispheric functional connectivity in patients with anisometropic and strabismic amblyopia: a resting-state fMRI study. Neuroradiology 2017; 59:517–524.

18.

Wang

Zhang

, et al. Interhemispheric Functional and Structural Disconnection in Alzheimer's Disease: a Combined Resting-State fMRI and DTI Study. PloS ONE 2015; 10:e0126310.

19.

Yan

Wang

Zuo

, et al. DPABI: data Processing & Analysis for (Resting-State) Brain Imaging. Neuroinformatics 2016; 14:339–351.

20.

Van Dijk

Sabuncu

Buckner

RL.

The influence of head motion on intrinsic functional connectivity MRI. Neuroimage 2012; 59:431–438.

21.

Goto

Abe

Aoki

, et al. Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra provides reduced effect of scanner for cortex volumetry with atlas-based method in healthy subjects. Neuroradiology 2013; 55:869–875.

22.

Di Pietro

McAuley

Parkitny

, et al. Primary motor cortex function in complex regional pain syndrome: a systematic review and meta-analysis. J Pain 2013; 14:1270–1288.

23.

Nash

Macefield

Klineberg

, et al. Changes in human primary motor cortex activity during acute cutaneous and muscle orofacial pain. J Orofac Pain 2010; 24:379–390.

24.

Zou

Lan

Wang

, et al. The potential roles of aquaporin 4 in amyotrophic lateral sclerosis. Neurol Sci 2019; 40:1541–1549.

25.

Geschwind

The organization of language and the brain. Science 1970; 170:940–944.

26.

Binder

Desai

Graves

, et al. Where is the semantic system? A critical review and meta-analysis of 120 functional neuroimaging studies. Cereb Cortex 2009; 19:2767–2796.

27.

Ploner

Schmitz

Freund

, et al. Differential organization of touch and pain in human primary somatosensory cortex. J Neurophysiol 2000; 83:1770–1776.

28.

Gazzola

Spezio

Etzel

, et al. Primary somatosensory cortex discriminates affective significance in social touch. Proc Natl Acad Sci USA 2012; 109:1657–1666.

29.

Longo

Pernigo

Haggard

Vision of the body modulates processing in primary somatosensory cortex. Neurosci Lett 2011; 489:159–163.

30.

Taylor-Clarke

Kennett

Haggard

Vision modulates somatosensory cortical processing. Curr Biol 2002; 12:233–236.

31.

Qin

Xuan

Liu

, et al. Functional Connectivity Density in Congenitally and Late Blind Subjects. Cereb Cortex 2015; 25:2507–2516.

32.

Kato

Takai

Hayashi

, et al. Expression and localization of aquaporin-4 in sensory ganglia. Biochem Biophys Res Commun 2014; 451:562–567.