Introduction

Diabetic retinopathy (DR) is a serious complication of long-term diabetes mellitus (DM),1 is primarily characterized by retinal microangiomas, retinal hemorrhage, macular edema, and retinal neovascularization.2 Diabetic retinopathy, as the primary microvascular complication of diabetes mellitus, stands as the leading cause of visual impairment in patients with diabetes mellitus.3 It is strongly linked to prolonged diabetes duration and inadequate control of glycemic and blood pressure levels.4 With the prolonged duration of diabetes, around 50% of patients develop diabetic retinopathy within 10 years of diabetes diagnosis, and this percentage may increase to 90% after 25 years.5 Epidemiological projections indicate that by 2030, the global population affected by diabetic retinopathy will escalate to 191 million individuals.6 Previous neuroimaging studies have demonstrated functional and structural alterations in the brains of individuals with diabetes.7,8 Furthermore, these functional alterations in individuals with diabetic retinopathy may contribute to the advancement of visual impairment.9 Currently, the precise neurophysiological mechanisms underlying visual loss and cognitive decline in patients with diabetic retinopathy remain elusive. Recent studies using amplitude of low-frequency fluctuations of resting-state BOLD imaging have revealed abnormal spontaneous brain activity across various brain regions in individuals with DR.10 An extensive study conducted over multiple years revealed a correlation between the severity of retinopathy in diabetic patients and a reduction in grey matter volume in specific brain regions, namely the right inferior frontal gyrus and right occipital lobe.11 There is a suggestion that the functional and structural alterations in the brains.9 Individuals with DR are at a heightened risk of developing neurodegenerative diseases.12 In a separate study, it was found that patients with proliferative diabetic retinopathy exhibited elevated apparent diffusion coefficient values in areas.13 A corresponding fMRI-based investigation demonstrated disrupted network module balance in the brains of patients with diabetic retinopathy, including alterations in the nodal centers of the default mode network (DMN) and the visual network (VN), alongside abnormal functional connectivity among the VN, DMN, salience network (SN), and sensorimotor network (SMN).14 Previous studies have illustrated the existence of atypical alterations in brain networks among DR patients. However, further research is warranted to enhance the depth of insight in this field.

The modular structure serves as one of the structural foundations for brain regions to function during typical resting brain activity.15 Each module within the brain network carries out distinct and autonomous cognitive functions, establishing connections between different modules when a task necessitates the collective involvement of multiple cognitive functions.16 Within the modular structure of the network, interconnections between modules are sparse, yet intra-module connections are dense, thereby preserving a balance between functional specialization and module integration, thereby creating a modular isolation model.17 Modular isolation models are currently applied in a range of biomedically significant studies. For instance, research has revealed a significant age-related increase in intramodular connectivity within the Default Mode Network (DMN), the Frontal-Parietal Network (FPN), and the Sensorimotor Network (SMN).18 The presence of decreased module separation within the Frontal-Parietal Network (FPN) has been observed in patients with major depressive disorder (MDD).17 Elevated modular segregation of the Frontal-Parietal Network (FPN), the Cingulo-Opercular Network (CON), and the cerebellum (Cere) has been observed in female patients with bulimia nervosa.19 At present, the neurophysiological mechanism behind the symptoms of patients with DR remains unclear, and the utilization of fMRI in investigating brain activity and network modifications in DR patients is limited. The modular segregation of whole brain functional networks in DR patients is currently unexplored. Considering these gaps, the potential utilization of modular isolation models for studying brain activity and network changes in DR patients is indicated.

Graph theory offers a distinctive quantitative methodology for delineating the organization of entire brain networks, especially through resting-state data20 and is currently an effective method for describing the segregation of brain network modules. The PC is a measure that quantifies both inter- and intra-module connections in brain networks, enabling the assessment of modular segregation in brain networks.21 Thus, with the assistance of a potent graph-theoretic metric like PC, we are able to uncover the anomalous modular segregation of brain networks in patients with DR.

In this study, we initially computed the mean PC value for each module and subsequently examined the number of intra- and inter-module connections to investigate which connections were linked to clinical symptoms of DR. Drawing from prior research, our hypotheses were as follows: (1) alterations in specific brain network module separations would be observed in the DR group compared to theHC group; (2) changes in inter-module connections would be evident in the DR group compared to the HC group.

Materials and Methods Participants

In the current study, a total of 89 participants were included, comprising 46 DR patients (22 males and 24 females) and 43 HC (22 males and 21 females), matched in terms of gender, age, and handedness. The participants were recruited from the Department of Ophthalmology at Jiangxi Provincial People’s Hospital. The number of ethical is 2023A0138.

All patients met the diagnostic criteria for diabetic retinopathy as per the International Clinical Diabetic Retinopathy Disease Severity Scale.22 The diagnostic criteria for DR patients incorporated fasting plasma glucose levels ≥ 7 mmol/L, random plasma glucose levels ≥ 11.1 mmol/L, or a 2-hour glucose level ≥ 11.1 mmol/L. Moreover, patients with nonproliferative DR exhibited features such as microaneurysms, hard exudates, and retinal hemorrhages. In our study, All patients with DR underwent fundus photography and fluorescein angiography.

The exclusion criteria for DR patients encompassed the presence of proliferative DR with retinal detachment, vitreous hemorrhage, additional ocular-related complications (such as cataract, glaucoma, high myopia, or optic neuritis), as well as the existence of diabetic nephropathy or diabetic neuropathy.

All HC met the specified criteria, inclusive of normal fasting plasma glucose levels are between 3.9 mmol/L (70 mg/dL) and 5.6 mmol/L (100 mg/dL), random plasma glucose levels < 11.1 mmol/L, HbA1c levels < 6.5%, absence of ocular diseases (myopia, cataracts, glaucoma, optic neuritis, or retinal degeneration), and binocular visual acuity ≥ 1.0.

Statement of Ethics

The study was conducted in compliance with the principles outlined in the Declaration of Helsinki and obtained approval from the ethical committee of Jiangxi Provincial People’s Hospital. Prior to participation, all individuals involved in the study provided written informed consent.

Image Acquisition

All participants were instructed to relax, keep their eyes closed, and maintain a motionless state without falling asleep during the resting-state fMRI scanning. Resting-state fMRI data were obtained using a 3T MRI system. Functional images were captured using a gradient-echo-planar imaging sequence with parameters including TR/TE = 2000 ms/25 ms, thickness = 3.0 mm, gap = 1.2 mm, acquisition matrix = 64 × 64, flip angle = 90°, field of view = 240 mm × 240 mm, voxel size = 3.6 mm × 3.6 mm × 3.6 mm, and 35 axial slices.

fMRI Preprocessing

The preprocessing was conducted using the Data Processing and Analysis Toolbox (DPABI) for Brain Imaging in MATLAB. This toolbox is specifically designed to minimize initial signal instability and confirm that participants had acclimated to the scanning environment.23 1) The initial 10 volumes for each participant were discarded, resulting in 230 remaining volumes. 2) Slice time correction was applied to the retained images. 3) Head motion correction was implemented to ensure all head movements were limited to translations of less than 2 mm or rotations of less than 2 degrees, with any exceedances leading to exclusion from further analyses. 4) Regressing out nuisance signals involved six motion parameters, the white matter signal, and the cerebrospinal fluid signal. 5) The spatial normalization was carried out into the Montreal Neurological Institute (MNI) template. 6) Detrending was performed to eliminate linear trends. 7) Temporal filtering (0.01–0.1 Hz) was applied to eliminate low-frequency drift and high-frequency noise.

Brain Network Construction and Graph Theory Analysis

All graph theory analyses were conducted using GRETNA, with a functional brain template subdividing the brain into 160 regions encompassing cerebral and cerebellar areas. In this study, the 160 regions of interest (ROIs) were classified into six functional modules identified as the default-mode network (DMN), frontal-parietal network (FPN), cingulo-opercular network (CON), sensorimotor network (SMN), visual network (VN), and the cerebellum (Cere).24

First, we calculated the PC to quantify the degree of module segregation. PC is calculated using the formula . In this formula, M is the set of modules, m refers to a module in a set of modules M, kim is the number of connections between node i and other nodes in the module m, and ki is the total number of connections of node i in the whole brain network.21 The PC quantifies the patterns of inter- and intra-module connectivity of node.

Furthermore, we computed the count of intra-module connections within each module and the count of inter-module connections in each pair of modules. The tally of inter-module connections is obtained by adding up the connections between all nodes in one module and all nodes in another module. Lastly, to investigate abnormal modular segregation at the node level, we determined the PC of each node within the module demonstrating atypical modular segregation. Figure 1 depicts a schematic representation of the processing steps.

Figure 1 Schematic representation of the processing steps. The entire brain is divided into six functional modules, consisting of the Default Mode Network (DMN), Frontal-Parietal Network (FPN), Cingulo-Opercular Network (CON), Sensorimotor Network (SMN), Visual Network (VN), and Cerebellum (Cere) based on prior studies.

Statistical Analyses

Both the chi-square (χ2) test and independent-samples t-test were utilized to compare clinical variables between the two groups using SPSS version 16.0 (SPSS Inc., Chicago, IL, USA).

A two-sample t-test was conducted to compare the between-group variances in the mean PC of each module, the counts of intra-module and inter-module connections, and the PC of each node within the module exhibiting abnormal modular segregation. To address multiple comparisons, a Bonferroni correction was applied with a significance threshold set at p < 0.05.

Results

The clinical characteristics of the participants were depicted in Table 1. Age and gender were comparable between the two groups, showing no significant differences (p > 0.05).

Table 1 Clinical-Demographic Characteristics of the DR and HC Groups

Different Mean PC of the Modules

As shown in Figure 2 and Table 2, the mean PC of DR group showed significantly lower on the FPN (t = −3.052, Pcorrected = 0.0030), the SMN (t = −3.774, Pcorrected = 0.0003) and the VN (t = −3.049, Pcorrected = 0.0030) than the HC group.

Table 2 Significant Differences in the Mean PC of Modules and the Inter-Module Connections

Figure 2 Between-group differences of mean participant coefficient (PC). Diabetic retinopathy (DR) patients showed significantly lower PC on the fronto-parietal network (FPN), the sensorimotor network (SMN) and the visual network (VN) than the healthy controls (HC).

Note: *Indicates p < 0.05.

Different Inter-Module Connections

Based on the above results, we further measured whether the effects of intra- and inter-module connectivity were anomalous. As shown in Figure 3 and Table 2, DR displayed significantly decreased inter-module connections between the DMN and SMN (t = 3.238, p uncorrected = 0.0017), FPN and VN (t = 3.653, p uncorrected = 0.0004) relative to the HC group.

Figure 3 Between-group comparison of inter-module connections and their correlation with HAMD scores. Compared with the healthy controls(HC), the Diabetic retinopathy (DR) patients exhibited significantly decreased inter-module connections between the default-mode network (DMN) and the sensorimotor network (SMN), the frontal-parietal network (FPN) and the visual network (VN).

Note: *indicates p < 0.01.

Different PC of Nodes

As shown in Figure 4 and Table 3, DR patients exhibited significantly changed PC in the left thalamus (t = 4.245, Pcorrected = 0.0001), the right frontal (t = 4.869, Pcorrected = 0.0001) and the right precentral gyrus (t = 4.212, Pcorrected = 0.0001).

Table 3 Significant Differences in the PC of Nodes

Figure 4 Significant differences in the PC of nodes.DR patients exhibited significantly changed PC in the left thalamus (t = 4.245, Pcorrected = 0.0001), the right frontal (t = 4.869, Pcorrected = 0.0001) and the right precentral gyrus (t = 4.212, Pcorrected = 0.0001).

Discussion

This study utilized graph theoretical analysis to investigate potential differences in resting-state network integration/segregation between diabetic retinopathy (DR) patients and healthy participants. Our results indicated that the DR group displayed significantly lower mean PC in the Frontal-Parietal Network (FPN), the Sensorimotor Network (SMN), and the Visual Network (VN) compared to the control group. Moreover, the DR group demonstrated enhanced inter-module connections between the Default Mode Network (DMN) and the SMN, and between the FPN and the VN. These outcomes suggest that, from a modularity standpoint, the onset of diabetic retinopathy is associated with irregularities in the intra- and inter-module connectivity of brain network modules.

We observed that DR patients exhibited a significantly reduced PC in the Frontal-Parietal Network (FPN) compared to the HC group, indicating increased intra-modular connections and decreased inter-modular connections in DR. The key brain regions of the FPN include the lateral prefrontal cortex (lPFC), midcingulate gyrus, and inferior parietal lobule (IPL). In a study conducted by Zhang et al, it was noted that patients with type 2 diabetes mellitus (T2DM) demonstrated notably lower functional connectivity (FC) between the pregenual anterior cingulate cortex (pACC) and the bilateral hippocampus, increased FC between the pACC and the bilateral lateral prefrontal cortex (LPFC) and left precentral gyrus, as well as decreased FC between the retrosplenial cortex (RSC) and right cerebellar Crus I.13 A prior study also indicated that individuals with poorer glycemic control exhibited reduced responses in the left temporal and prefrontal regions during encoding.25 Matthew et al discovered that the frontoparietal control network (FPCN) has a pivotal function in executive control, with its subsystem, the FPCN, potentially specializing in the management of visuospatial perceptual attention.26 The Sensorimotor Network (SMN) is predominantly composed of sensory regions (e.g, posterior central gyrus) and motor areas (eg, anterior central gyrus), with extensions into supplementary motor regions. In a previous study, Frøkjær et al identified decreased cortical thickness in the posterior central gyrus of the brain in individuals with type 1 diabetes mellitus.27 Several years later, abnormal activity in central cortical areas among patients with diabetes mellitus (DM) during the resting state was once more detected through resting-state EEG frequency analysis and source localization.28 In conclusion, the findings of this study suggest that patients with DR exhibit various abnormalities in brain networks, specifically within the Frontal-Parietal Network (FPN) and Sensorimotor Network (SMN), which could potentially signify a pathogenic mechanism of DR.

The Visual Network (VN) is situated in the occipital and temporal lobes of the brain, primarily responsible for the processing of visual information.29 One of the primary clinical symptoms in DR patients is severe vision loss. It has been commonly reported in prior studies that diabetes is linked with both functional and structural alterations in the visual center. In our study, we verified that the mean Participation Coefficient (PC) of the Visual Network (VN) was significantly lower in the DR group compared to the healthy control (HC) group. Previous research has indicated that individuals with Type 2 Diabetes Mellitus (T2DM) often exhibit reduced grey matter volume in the middle occipital gyrus.30 Another study was shown that individuals with Type 2 Diabetes Mellitus (T2DM) without retinopathy exhibit reduced grey matter volume in the occipital lobe.31 Qi et al reported that DR patients exhibited FC disruptions between V1 and higher visual regions at rest, which may reflect the aberrant information communication in the V1 area of DR individuals.32 This indicates that the severe vision loss symptoms observed in patients with DRcould be attributed to diabetes-induced alterations in the brain’s visual network.

Moreover, compared to the healthy controls (HC), the patients with diabetic retinopathy (DR) displayed significantly reduced inter-module connections between the Default Mode Network (DMN) and the Sensorimotor Network (SMN), as well as between the Frontal-Parietal Network (FPN) and the Visual Network (VN). In a study by Chen et al, it was reported that individuals with Type 2 Diabetes Mellitus (T2DM) present with increased path lengths, decreased overall efficiency, and disrupted long-distance connections in functional brain networks.33 The graph-theoretic analysis indicates that the heightened module separation of the Frontal-Parietal Network (FPN), the Sensorimotor Network (SMN), and the Visual Network (VN) might be associated with the reduced inter-module connections between the Default Mode Network (DMN) and the SMN, and between the FPN and the VN. This novel insight offers a new perspective on investigating the neurobiological mechanisms underlying DR. This study partly indicates a decline in the information transmission efficiency of specific brain functional regions in individuals with diabetic retinopathy. In future treatments and rehabilitation for individuals with diabetic retinopathy, there should be a stronger focus on restoring cognitive function, sensory-motor skills, and visual abilities specific to DR patients.

At the node level, notable alterations were observed in the PC of the thalamus, frontal gyrus, and precentral gyrus within the DR group in comparison to the HC group. The preceding discussion has highlighted the existence of brain network irregularities in the frontal and central gyrus among individuals with DR. Furthermore, in a study by Chen et al, it was documented that the functional connectivity of the thalamus is notably decreased in patients with Type 2 Diabetes Mellitus (T2DM) relative to the HC.34 In their study, Wang et al identified elevated resting-state functional connectivity (rsFC) in cortical regions across various subdivisions of the thalamus in individuals with Type 2 Diabetes Mellitus (T2DM).35 Liu et al proposed that there is an elevation in thalamo-cortical functional connectivity in individuals with Type 2 Diabetes Mellitus (T2DM) as revealed by resting-state functional MRI.36 In summary, there are also aberrant alterations in the thalamus within the brain network of DR. In our current study, several limitations should be acknowledged. Firstly, (1) the PC calculations might be influenced by the number of nodes in the module. Secondly, (2) the study’s small sample size and brief resting-state fMRI acquisition time necessitate a larger sample size and longer scan duration to enhance result accuracy. Thirdly, (3) the participants in this study were geographically limited and did not represent other ethnic groups. Lastly, (4) only Dosenbach’s template was utilized for network construction, suggesting that a different template should be considered to validate the study results in future investigations.(5) Diabetic retinopathy may be linked to a range of systemic conditions such as circadian rhythm and hypertension. These concurrent conditions may influence brain networks, and upcoming studies aim to mitigate these confounding variables.

Conclusions

This study revealed heightened modular segregation in the FPN, SMN, and VN, along with reduced inter-modular connectivity between the DMNand SMN, as well as between the FPN and VN. Furthermore, significant alterations in the PCwere observed in the thalamus, frontal gyrus, and precentral gyrus. These results indicate a potential association between the pathogenesis of diabetic retinopathy and irregularities in intra- and inter-modular connectivity within brain networks from a modular perspective. In future studies, we will integrate multimodal magnetic resonance imaging for analyzing the structural networks related to diabetic retinopathy. Concurrently, we will investigate pre- and post-treatment alterations in brain networks among DR patients to pinpoint sensitive neurobiological markers that can serve as crucial imaging references for future diabetic retinopathy treatments.

Data Sharing Statement

The raw data supporting the conclusions of this article will be made available. Further inquiries can be directed to the corresponding author.

Acknowledgments

We acknowledge the assistance provided by the Natural Science Foundation of Jiangxi Province (20212BAB216058), Jiangxi Provincial Health Technology Project (202210012, 202310114 and 202410008), and Jiangxi Provincial traditional Chinese Technology Project (2022B840 and 2023A0138).

Disclosure

The authors declare that they have no conflicts of interest with regard to this work.

References

1. Agarwal P, Jindal A, Saini VK, Jindal S. Advances in diabetic retinopathy. Indian J Endocrinol Metab. 2014;18(6):772–777. PMID: 25364670; PMCID: PMC4192980. doi:10.4103/2230-8210.140225

2. Biswas S, Sarabusky M, Chakrabarti S. Diabetic retinopathy, lncRNAs, and Inflammation: a dynamic, interconnected network. J Clin Med. 2019;8(7):1033. PMID: 31337130; PMCID: PMC6678747. doi:10.3390/jcm8071033

3. Yi S, Yufeng X, Xiling L, et al. Burden of vision loss due to diabetic retinopathy in China from 1990 to 2017: findings from the global burden of disease study.[J]. Acta ophthalmologica. 2020;99(2):e267–e273. doi:10.1111/aos.14573

4. Yau JW, Rogers SL, Kawasaki R, et al. Global prevalence and major risk factors of diabetic retinopathy. Diabetes Care. 2012;35(3):556–564. PMID: 22301125; PMCID: PMC3322721. doi:10.2337/dc11-1909

5. Calderon GD, Juarez OH, Hernandez GE, Punzo SM, De la Cruz ZD. Oxidative stress and diabetic retinopathy: development and treatment. Eye. 2017;31(8):1122–1130. PMID: 28452994; PMCID: PMC5558229. doi:10.1038/eye.2017.64

6. Ting DS, Cheung GC, Wong TY. Diabetic retinopathy: global prevalence, major risk factors, screening practices and public health challenges: a review. Clin Exp Ophthalmol. 2016;44(4):260–277. PMID: 26716602. doi:10.1111/ceo.12696

7. Chen Z, Li L, Sun J, Ma L. Mapping the brain in type II diabetes: voxel-based morphometry using DARTEL. Eur J Radiol. 2012;81(8):1870–1876. PMID: 21546180. doi:10.1016/j.ejrad.2011.04.025

8. van Duinkerken E, Schoonheim MM, Sanz-Arigita EJ, et al. Resting-state brain networks in type 1 diabetic patients with and without microangiopathy and their relation to cognitive functions and disease variables. Diabetes. 2012;61(7):1814–1821. PMID: 22438575; PMCID: PMC3379683. doi:10.2337/db11-1358

9. Li YM, Zhou HM, Xu XY, Shi HS. Research progress in mri of the visual pathway in diabetic retinopathy. Curr Med Sci. 2018;38(6):968–975. PMID: 30536057. doi:10.1007/s11596-018-1971-5

10. Wang ZL, Zou L, Lu ZW, et al. Abnormal spontaneous brain activity in type 2 diabetic retinopathy revealed by amplitude of low-frequency fluctuations: a resting-state fMRI study. Clin Radiol. 2017;72(4):340.e1–340.e7. PMID: 28041652. doi:10.1016/j.crad.2016.11.012

11. Hugenschmidt CE, Lovato JF, Ambrosius WT, et al. The cross-sectional and longitudinal associations of diabetic retinopathy with cognitive function and brain MRI findings: the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. Diabetes Care. 2014;37(12):3244–3252. PMID: 25193529; PMCID: PMC4237980. doi:10.2337/dc14-0502

12. Ho WK, Katherine H, Cecilia P, et al. Diabetic retinopathy and risk of stroke: a secondary analysis of the ACCORD eye study.[J]. Stroke. 2020;51(12):STROKEAHA120030350.

13. Zhang D, Huang Y, Liu S, et al. Structural and functional connectivity alteration patterns of the cingulate gyrus in Type 2 diabetes. Ann Clin Transl Neurol. 2023;10(12):2305–2315. PMID: 37822294; PMCID: PMC10723245. doi:10.1002/acn3.51918

14. Huang X, Tong Y, Qi CX, Xu YT, Dan HD, Shen Y. Disrupted topological organization of human brain connectome in diabetic retinopathy patients. Neuropsychiatr Dis Treat. 2019;15:2487–2502. PMID: 31695385; PMCID: PMC6717727. doi:10.2147/NDT.S214325

15. Valencia M, Pastor MA, Fernández-Seara MA, Artieda J, Martinerie J, Chavez M. Complex modular structure of large-scale brain networks. Chaos. 2009;19(2):023119. PMID: 19566254. doi:10.1063/1.3129783

16. Bertolero MA, Yeo BT, D’Esposito M. The modular and integrative functional architecture of the human brain. Proc Natl Acad Sci. 2015;112(49):E6798–807. PMID: 26598686; PMCID: PMC4679040. doi:10.1073/pnas.1510619112

17. Zhihui L, Wei Z, Donglin W, et al. Decreased modular segregation of the frontal–parietal network in major depressive disorder#13;[J]. Frontiers in Psychiatry. 2022;13:929812. doi:10.3389/fpsyt.2022.929812

18. Lan Z, Zhu LL, Wu YK, et al. Aberrant modular segregation of brain networks in female patients with bulimia nervosa. Int J Eat Disord. 2023;56(7):1353–1364. PMID: 36951235. doi:10.1002/eat.23939

19. Wang Z, Lu Z, Li J, et al. Evaluation of apparent diffusion coefficient measurements of brain injury in type 2 diabetics with retinopathy by diffusion-weighted MRI at 3.0 T. Neuroreport. 2017;28(2):69–74. PMID: 27846040. doi:10.1097/WNR.0000000000000703

20. Davis FC, Knodt AR, Sporns O, et al. Impulsivity and the modular organization of resting-state neural networks. Cereb Cortex. 2013;23(6):1444–1452. PMID: 22645253; PMCID: PMC3643719. doi:10.1093/cercor/bhs126

21. Guimerà R, Nunes Amaral LA. Functional cartography of complex metabolic networks. Nature. 2005;433(7028):895–900. PMID: 15729348; PMCID: PMC2175124. doi:10.1038/nature03288

22. Wilkinson C, Ferris FL, Klein RE, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales[J]. Ophthalmology. 2003;110(9):1677–1682. doi:10.1016/S0161-6420(03)00475-5

23. Yan CG, Wang XD, Zuo XN, Zang YF. DPABI: data processing & analysis for (Resting-State) brain imaging. Neuroinformatics. 2016;14(3):339–351. PMID: 27075850. doi:10.1007/s12021-016-9299-4

24. Dosenbach NFU, Nardos B, Cohen AL, et al. Prediction of individual brain maturity using fMRI[J]. Science. 2010;329(5997):1358–1361. doi:10.1126/science.1194144

25. Embury CM, Lord GH, Drincic AT, Desouza CV, Wilson TW. Glycemic control level alters working memory neural dynamics in adults with type 2 diabetes. Cereb Cortex. 2023;33(13):8333–8341. PMID: 37005060; PMCID: PMC10321117. doi:10.1093/cercor/bhad119

26. Dixon ML, De La Vega A, Mills C, et al. Heterogeneity within the frontoparietal control network and its relationship to the default and dorsal attention networks. Proc Natl Acad Sci. 2018;115(7):E1598–E1607. PMID: 29382744; PMCID: PMC5816169. doi:10.1073/pnas.1715766115

27. Frøkjær JB, Brock C, Søfteland E, et al. Macrostructural brain changes in patients with longstanding type 1 diabetes mellitus - a cortical thickness analysis study. Exp Clin Endocrinol Diabetes. 2013;121(6):354–360. PMID: 23757052. doi:10.1055/s-0033-1345120

28. Frøkjær JB, Graversen C, Brock C, et al. Integrity of central nervous function in diabetes mellitus assessed by resting state EEG frequency analysis and source localization[J]. J diabet complicat. 2016;31(2):400–406. doi:10.1016/j.jdiacomp.2016.11.003

29. Wang K, Jiang T, Yu C, et al. Spontaneous activity associated with primary visual cortex: a resting-state FMRI study. Cereb Cortex. 2008;18(3):697–704. PMID: 17602140. doi:10.1093/cercor/bhm105

30. Zhang Y, Zhang X, Zhang J, et al. Gray matter volume abnormalities in type 2 diabetes mellitus with and without mild cognitive impairment. Neurosci Lett. 2014;562:1–6. PMID: 24434688. doi:10.1016/j.neulet.2014.01.006

31. Ferreira FS, Pereira JMS, Reis A, et al. Early visual cortical structural changes in diabetic patients without diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol. 2017;255(11):2113–2118. PMID: 28779362. doi:10.1007/s00417-017-3752-4

32. Qi CX, Huang X, Tong Y, Shen Y. Altered functional connectivity strength of primary visual cortex in subjects with diabetic retinopathy. Diabetes Metab Syndr Obes. 2021;13(14):3209–3219. doi:10.2147/DMSO.S311009.eCollection

33. Chen GQ, Zhang X, Xing Y, Wen D, Cui GB, Han Y. Resting-state functional magnetic resonance imaging shows altered brain network topology in Type 2 diabetic patients without cognitive impairment. Oncotarget. 2017;8(61):104560–104570. PMID: 29262661; PMCID: PMC5732827. doi:10.18632/oncotarget.21282

34. Chen YC, Xia W, Qian C, Ding J, Ju S, Teng GJ. Thalamic resting-state functional connectivity: disruption in patients with type 2 diabetes. Metab Brain Dis. 2015;30(5):1227–1236. PMID: 26116166. doi:10.1007/s11011-015-9700-2

35. Jie W, Shanlei Z, Datong D, et al. Compensatory thalamocortical functional hyperconnectivity in type 2 Diabetes Mellitus[J]. Brain Imag & Behav. 2022;16(6):2556–2568. doi:10.1007/s11682-022-00710-0

36. Liu X, Xu X, Mao C, et al. Increased thalamo-cortical functional connectivity in patients with diabetic painful neuropathy: a resting-state functional MRI study. Exp Ther Med. 2021;21(5):509. PMID: 33791018; PMCID: PMC8005696. doi:10.3892/etm.2021.9940