The aim of this study was to investigate the interactive effect between aging and type 2 diabetes mellitus (T2DM) on brain glucose metabolism, individual metabolic connectivity, and network properties. Using a 2 × 2 factorial design, 83 patients with T2DM (40 elderly and 43 middle-aged) and 69 sex-matched healthy control subjects (HCs) (34 elderly and 35 middle-aged) underwent 18F-fluorodeoxyglucose positron emission tomography/magnetic resonance scanning. Jensen-Shannon divergence was applied to construct individual metabolic connectivity and networks. The topological properties of the networks were quantified using graph theoretical analysis. The general linear model was used to mainly estimate the interaction effect between aging and T2DM on glucose metabolism, metabolic connectivity, and network. There was an interaction effect between aging and T2DM on glucose metabolism, metabolic connectivity, and regional metabolic network properties (all P < 0.05). The post hoc analyses showed that compared with elderly HCs and middle-aged patients with T2DM, elderly patients with T2DM had decreased glucose metabolism, increased metabolic connectivity, and regional metabolic network properties in cognition-related brain regions (all P < 0.05). Age and fasting plasma glucose had negative correlations with glucose metabolism and positive correlations with metabolic connectivity. Elderly patients with T2DM had glucose hypometabolism, strengthened functional integration, and increased efficiency of information communication mainly located in cognition-related brain regions. Metabolic connectivity pattern changes might be compensatory changes for glucose hypometabolism.

With improvement of living standards and medical levels, the life span of human beings has been greatly extended so that the number of elderly people has also increased. Because of long-term survival, elderly people commonly live with chronic disease; therefore, the management of chronic diseases has become an important issue in an aging society (1). Type 2 diabetes mellitus (T2DM), characterized by insulin resistance and hyperglycemia, is a common chronic metabolic disease, and its complications, comorbidities, and related functional disabilities have been great challenges for elderly people (24). It is estimated that the total economic costs of T2DM and its complications have risen by 26% from 2012 to 2017, mainly among elderly people (5,6). Because both the central and peripheral nervous systems can be affected in T2DM and aging, the related impact on the brain is of great concern. Therefore, investigations of the aging-dependent and T2DM-related changes of the brain are important for early interventions for potential brain abnormality.

Previous studies have focused on the main effect of aging or T2DM. Compared with elderly people without T2DM or younger patients with T2DM, elderly patients with T2DM showed poorer performance in cognitive function tests and more accelerated cognitive function decline (710). Vascular inflammation and brain atrophy might be the reasons why elderly patients with T2DM are more susceptible to cerebrovascular damages, leading to serious complications, including vascular cognitive impairment and Alzheimer disease (7,1115). Knowledge of aging-dependent and T2DM-related brain changes is still limited.

Because the human brain usually works as a complex network, it is essential to investigate specific changes in the brain with regard to its whole network. Compared with blood-oxygen-level-dependent–based analysis in functional MRI (fMRI), positron emission tomography (PET) can detect brain glucose metabolism and, thus, a more direct state of local neuronal activities (16,17). Different from an individual network based on signal time series in fMRI, only one average brain metabolic map can be acquired for one individual in PET scanning. Therefore, a group-level metabolic brain network was constructed in most previous studies, which could only describe the metabolic covariance and topological properties in a group of individuals (18,19). In that way, specific information within an individual brain may be lost (20). Recently, several novel distribution-divergence–based methods, including Kullback-Leibler divergence similarity estimation (KLSE), Jensen-Shannon divergence (JSD), and exponential function of multivariate Euclidean distance, have been used in construction of the individual metabolic network by quantifying the similarity of the probability density functions (PDFs) between any pair of regions of interest (ROIs) (2023). It has been proven that the distribution-divergence–based methods are reliable in mapping the individual brain networks (2426). Because the interindividual variability has been largely preserved, the individual brain networks perform better in detecting more subtle and specific brain changes in certain diseases or pathological conditions than group-level brain networks (20). As described above, given the important impact of T2DM on elderly patients, it might be insufficient to investigate the main effect of aging or T2DM (2729). That is, it will be of great value to explore the principle of abnormality within an individual brain under both conditions, which might provide potential therapeutic targets to simultaneously prevent or postpone the impact of aging and T2DM on the brain. To our knowledge, few studies have focused on the superimposed impact between aging and T2DM on brain glucose metabolism, especially the individual metabolic connectivity and network.

In the current study, we investigated the interaction effect between aging and T2DM on brain metabolic changes from three different aspects simultaneously, including local standardized uptake value (SUV), metabolic connectivity, and individual metabolic network constructed with JSD, a statistical method used to compare the interregional similarity of SUV distribution. The findings might provide new insights and information for understanding brain changes in elderly patients with T2DM from a different perspective.

Study Design

We used a 2 × 2 factorial design to investigate the interaction effect between aging and T2DM on brain glucose metabolism and the individual metabolic connectivity and network. Because each factor included two levels (aging: elderly, 60–80 years of age, and middle-aged, 40–59 years of age; diabetes status: T2DM and healthy control subjects [HCs]), participants were assigned to four groups: 1) elderly patients with T2DM; 2) elderly HCs; 3) middle-aged patients with T2DM; and 4) middle-aged HCs (30). The study protocol was approved by the Ethics Committee of Yueyang Hospital of Integrated Traditional Chinese and Western Medicine affiliated with Shanghai University of Traditional Chinese Medicine (No. 2020-188). Written informed consent was provided by all participants before the study started.

Participants

Participants who received 18F-fluorodeoxyglucose PET/magnetic resonance (18F-FDG-PET/MR) scanning between January 2017 and December 2019 were recruited for analysis. The inclusion criteria for participants with T2DM included 1) aged 40–80 years, 2) diagnosis of T2DM (A1C ≥48 mmol/mol, fasting plasma glucose [FPG] ≥7.0 mmol/L, 2-h plasma glucose ≥11.11 mmol/L during a 75-g oral glucose tolerance test, or classic symptoms of hyperglycemia or hyperglycemic crisis with a random plasma glucose ≥11.1 mmol/L) (31), and 3) BMI ≤28 kg/m2 (32). The inclusion criteria for HCs included 1) aged 40–80 years and 2) BMI ≤28 kg/m2. Exclusion criteria included 1) a history of confirmed stroke, head injury, Parkinson disease, major depression, or other neurological or psychiatric diseases and 2) contraindications of 18F-FDG-PET/MR scanning, such as cardiac pacemaker and heart stent implantation.

18F-FDG-PET/MR Scanning Image Acquisition and Preprocessing

The 18F-FDG-PET/MR scanning images were obtained in the supine position with a hybrid 3T Biograph mMR system (Siemens Healthineers, Erlangen, Germany). Before scanning, the participant was required to fast for >6 h, and plasma glucose was controlled to be <7 mmol/L. All participants then received an intravenous injection of a mean dose of 4.5 MBq/kg of 18F-FDG. During the 30-min uptake period, participants were asked to be awake but relaxed in a quiet room. PET/MR scanning was performed in five bed positions (3 min per bed position). Reconstruction was performed by the mode of the ordinary Poisson ordered subset expectation maximization algorithm with point spread function for three different iterations, 21 subsets, and 4-mm Gaussian filtering (33). The PET images were corrected for attenuation of annihilation radiation with the MRI-based attenuation correction (34).

The Digital Imaging and Communications in Medicine format of PET images was converted into the Neuroimaging Informatics Technology Initiative format with dcm2nii version 12 12 2012 (Chris Rorden, Columbia University, New York, NY) before preprocessing. The preprocessing procedures were performed using Statistical Parametric Mapping 8 software (https://www.fil.ion.ucl.ac.uk/spm/) implemented in MATLAB 2013b (MathWorks, Inc., Natick, MA). Briefly, preprocessing included the following approaches. First, images were manually reoriented and the anterior commissure reset to the origin of the three-dimensional Montreal Neurological Institute space. Second, all PET images were normalized to the Montreal Neurological Institute space through the standard PET template provided with the Statistical Parametric Mapping 8 software and then resampled to a resolution of 3 × 3 × 3 mm3 (35). As part of the quality control process, the normalized PET images were visually checked to ensure the correctness of registration. Finally, normalized PET images were smoothed with a Gaussian kernel of 10-mm full width at half maximum.

Glucose Metabolism

As described in previous studies, after smoothing, the intensity of the 18F-FDG uptake value of each voxel was normalized by the mean uptake value of the global brain to minimize the influence of the uptake differences across individuals, which is referred to as the SUV (36,37). The SUV represents the glucose metabolism, indicating the energy consumption requirements of a specific brain region. On the basis of the automated anatomical labeling (AAL) atlas, the SUV of the brain was parcellated into 90 ROIs and processed using ROI-wise analysis.

JSD-Based Metabolic Connectivity and Individual Metabolic Network Construction

Group-level metabolic network describes the brain metabolic covariation and network properties in a group of individuals, but specific information within an individual might get lost (20). In the current study, the metabolic connectivity and network of each individual were calculated and constructed using JSD (23). Briefly, the JSD was used to compare the probability distribution of the SUVs of a group of voxels in an ROI with that in another ROI in an individual brain. The approach provided the metabolic connectivity between any pair of brain regions in an individual brain for further statistical analyses in four groups of participants.

Construction of the individual metabolic connectivity and network mainly included four steps. First, the smoothed, global intensity–normalized PET images were parcellated into 90 ROIs according to the AAL atlas (38). Second, the SUV of all voxels within a given ROI were used to estimate the PDF of the ROI using the nonparametric density estimation kernel density estimator (39). The PDFs of all 90 ROIs were estimated. Third, the JSD was applied to compare the similarity of PDF distribution between every pair of ROIs by calculating the normalized symmetrical score on the basis of Eq. 1 and 2, referred to as JS (P ǁ Q) (40). The greater the similarity of two PDFs, the smaller JS (P ǁ Q) and the less information provided with PDFs of both ROIs.
formula
formula
where P and Q represent the PDFs of ROIs, the operator ǁ indicates divergence, and DKL is calculated as the KLSE. (4) The greater similarity of PDFs between two ROIs reflects the nearer consistent functional activity levels in these brain regions. Thus, we used the value of 1 − JS (P ǁ Q) to represent the strength of the metabolic connectivity between two ROIs. Therefore, the value of the strength of the metabolic connectivity between either pair of ROIs constituted a 90 × 90 metabolic correlation adjacency matrix Rij (i = 1, 2, 3, …, N; j = 1, 2, 3, …, N; N = 90) for each participant, which is referred to as the individual metabolic network.

Graph Theoretical Analysis

Graph theoretical analysis was performed according to our previous description (41). Briefly, on the basis of the individual metabolic network, topological properties were quantified with the GRETNA toolbox version 3 (https://www.nitrc.org/projects/gretna/) (42). Each ROI was regarded as a node of the metabolic network, while the metabolic connectivity between a pair of ROIs was regarded as their edge. Global network properties included the clustering coefficient, characteristic path length, σ, global efficiency, local efficiency, assortativity, synchronization, and hierarchy, while regional network properties included degree centrality, global efficiency, and local efficiency of a given node. The specific definitions and equations of these topological properties have been presented in previous studies (4346).

Statistical Analysis

The normality test was used to determine whether the data had a normal distribution. Then the normally distributed data were described as mean ± SD, and enumeration data were described as n (%). The baseline characteristics of the four study groups were analyzed with one-way ANOVA and χ2 test. The general linear model was used to estimate the interaction effect between aging and T2DM on glucose metabolism, metabolic connectivity, and the individual ROI-wise metabolic network properties. If there was an interaction effect between aging and T2DM, the simple effect of aging or T2DM was further analyzed by using post hoc analysis. The significance level was set at α = 0.05 (two-tailed), and Bonferroni method was applied for correction of multiple comparisons (corrected α, α′ ∼0.05/n, where n represents comparison times). SPSS version 21 software (IBM Corporation, Chicago, IL) was used for the statistical analysis.

Partial Correlation Analysis

Partial correlation analysis was performed with SPSS version 21 software. The correlations between either age or FPG and glucose metabolism, strength of metabolic connectivity, and individual metabolic network properties were explored while controlling the effect of the other factor (FPG or age). The correlations between BMI and glucose metabolism, strength of metabolic connectivity, and individual metabolic network properties were also explored while simultaneously controlling the effects of both age and FPG. The selection of ROIs was based on the brain region’s presented interaction effect between aging and T2DM in the general linear model analysis of glucose metabolism, metabolic connectivity, and individual metabolic network properties.

The results were visualized with GraphPad Prism 5 (GraphPad Software, San Diego, CA), MATLAB 2013b, Stata MP 16.0 (StataCorp LLC, College Station, TX), and Circos version 3 (Free Software Foundation, Boston, MA) (47). The processes of PET image preprocessing, metabolic connectivity calculation, and individual metabolic network construction of a participant are shown in Fig. 1.

Figure 1

Flow diagram of PET image preprocessing, metabolic connectivity calculation, and individual metabolic network construction for a participant.

Figure 1

Flow diagram of PET image preprocessing, metabolic connectivity calculation, and individual metabolic network construction for a participant.

Close modal

Data and Resource Availability

The data sets analyzed during this study are available from the corresponding author upon reasonable request.

Baseline Characteristics

A total of 152 participants were screened, including 83 with T2DM (40 elderly and 43 middle-aged) and 69 sex- and BMI-matched HCs (34 elderly and 35 middle-aged). After image quality control, 15 original PET images (3 from elderly patients with T2DM, 2 elderly HCs, 9 from middle-aged patients with T2DM, and 1 from a middle-aged HC) were excluded because of image artifacts, and no normalized PET image was excluded from further analysis (Supplementary Fig. 1). We finally enrolled 37 elderly patients with T2DM (65.70 ± 0.59 years), 32 elderly HCs (67.43 ± 1.57 years), 34 middle-aged patients with T2DM (51.29 ± 0.91 years), and 34 middle-aged HCs (49.38 ± 1.01 years) (Table 1 and Supplementary Table 1). The FPG of the elderly and middle-aged patients with T2DM were all >7 mmol/L with a mean ± SD of 9.16 ± 2.87 and 8.36 ± 1.91 mmol/L, respectively, and the FPG of elderly and middle-aged HCs were normal, with a mean ± SD of 5.38 ± 0.55 and 5.33 ± 0.62 mmol/L, respectively. There was no intergroup difference of BMI (F = 0.798, P = 0.497), sex (χ2 = 4.801, P = 0.187), hypertension (χ2 = 5.868, P = 0.118), or hyperlipidemia (χ2 = 1.287, P = 0.732).

Table 1

Baseline participant characteristics

CharacteristicElderly patients with T2DM
(n = 37)
Elderly HCs
(n = 32)
Middle-aged patients with T2DM
(n = 34)
Middle-aged HCs
(n = 34)
F2P
Age, years 65.70 ± 3.60 67.75 ± 5.03 51.29 ± 5.29 50.09 ± 5.21 127.24 <0.001 
FPG, mmol/L 9.16 ± 2.87 5.38 ± 0.55 8.36 ± 1.91 5.33 ± 0.62 41.700 <0.001 
BMI, kg/m2 24.87 ± 2.70 23.83 ± 3.09 25.00 ± 4.34 24.32 ± 3.65 0.798 0.497 
Sex       
 Male 27 (73) 23 (72) 31 (91) 27 (79) 4.801 0.187 
 Female 10 (27) 9 (28) 3 (9) 7 (21)   
Hypertension       
 Yes 20 (54) 16 (50) 12 (35) 10 (30) 5.868 0.118 
 No 17 (47) 16 (50) 22 (65) 24 (70)   
Hyperlipidemia       
 Yes 13 (35) 9 (28) 11 (32) 8 (24) 1.287 0.732 
 No 24 (65) 23 (72) 23 (68) 26 (76)   
CharacteristicElderly patients with T2DM
(n = 37)
Elderly HCs
(n = 32)
Middle-aged patients with T2DM
(n = 34)
Middle-aged HCs
(n = 34)
F2P
Age, years 65.70 ± 3.60 67.75 ± 5.03 51.29 ± 5.29 50.09 ± 5.21 127.24 <0.001 
FPG, mmol/L 9.16 ± 2.87 5.38 ± 0.55 8.36 ± 1.91 5.33 ± 0.62 41.700 <0.001 
BMI, kg/m2 24.87 ± 2.70 23.83 ± 3.09 25.00 ± 4.34 24.32 ± 3.65 0.798 0.497 
Sex       
 Male 27 (73) 23 (72) 31 (91) 27 (79) 4.801 0.187 
 Female 10 (27) 9 (28) 3 (9) 7 (21)   
Hypertension       
 Yes 20 (54) 16 (50) 12 (35) 10 (30) 5.868 0.118 
 No 17 (47) 16 (50) 22 (65) 24 (70)   
Hyperlipidemia       
 Yes 13 (35) 9 (28) 11 (32) 8 (24) 1.287 0.732 
 No 24 (65) 23 (72) 23 (68) 26 (76)   

Data are mean ± SD or n (%).

Decreased Glucose Metabolism in the Left Triangle of the Inferior Frontal Gyrus in Elderly Patients With T2DM

There was a significant interaction effect between aging and T2DM on the SUV in the left triangle of the inferior frontal gyrus (F = 7.330, P = 0.008) (Fig. 2A), showing a decrease in elderly patients with T2DM compared with elderly HCs (mean difference 0.31 [95% CI 0.19, 0.43]) and compared with middle-aged patients with T2DM (mean difference 0.26 [95% CI 0.14, 0.38]) (Supplementary Table 2 and Fig. 2B).

Figure 2

A: The interaction effect between aging and T2DM on the SUV in the left triangle of the inferior frontal gyrus. B: Comparison of SUV of the left triangle of the inferior frontal gyrus between two groups in post hoc analysis. *P < 0.001.

Figure 2

A: The interaction effect between aging and T2DM on the SUV in the left triangle of the inferior frontal gyrus. B: Comparison of SUV of the left triangle of the inferior frontal gyrus between two groups in post hoc analysis. *P < 0.001.

Close modal

Increased Strength of Metabolic Connectivity in Elderly Patients With T2DM

The left triangle of the inferior frontal gyrus was defined as a seed ROI to calculate its metabolic connectivity with the rest of the 89 ROIs according to the AAL atlas. Significant interaction effect between aging and T2DM on the strength of metabolic connections were observed in the left anterior cingulate cortex (ACC) (F = 4.949, P = 0.028), left (F = 7.326, P = 0.008) and right middle cingulate cortex (MCC) (F = 4.753, P = 0.031), left angular gyrus (F = 5.572, P = 0.020), and left (F = 11.018, P = 0.001) and right Heschl gyrus (F = 4.222, P = 0.042) (Table 2). Post hoc analyses showed strengthened metabolic connectivity between the left triangle of the inferior frontal gyrus and six brain regions (left ACC, bilateral MCCs, left angular gyrus, and bilateral Heschl gyri) in the elderly patients with T2DM compared with the elderly HCs and middle-aged patients with T2DM (Supplementary Tables 3–8 and Fig. 3). Abbreviations and full names of the 90 brain regions in the AAL atlas are listed in Supplementary Table 9.

Figure 3

Post hoc analyses of the strength of metabolic connectivity of the left triangle of the inferior frontal gyrus (IFGtri.L). Compared with the remaining three groups, metabolic connectivity between IFGtri.L and brain regions, including left ACC (ACC.L), left MCC (MCC.L), right MCC (MCC.R), left angular gyrus (ANG.L), left Heschl gyrus (HES.L), and right Heschl gyrus (HES.R) increased in the elderly patients with T2DM. AMYG, amygdala; CAL, calcarine; CAU, caudate; CUN, cuneus; FG, fusiform gyrus; HIP, hippocampus; INS, insula; IOG, inferior occipital gyrus; IPG, inferior parietal, but supramarginal and angular gyrus; ITG, inferior temporal gyrus; L, left; LING, lingual gyrus; med, medial; MFG, middle frontal gyrus; morb, medial orbital; MOG, middle occipital gyrus; MTG, middle temporal gyrus; OLF, olfactory cortex; oper, opercular part; orb, orbital part; p, temporal pole; PAL, pallidum; PCC, posterior cingulate cortex; PCL, paracentral lobule; PCUN, precuneus; PHIP, parahippocampal gyrus; PoCG, postcentral gyrus; PreCG, precental gyrus; PUT, putamen; R, right; REG, gyrus rectus; ROL, Rolandic operculum; SFG, superior frontal gyrus; SMA, supplementary motor area; SMG, supramarginal gyrus; SOG, superior occipital gyrus; SPG, superior parietal gyrus; STG, superior temporal gyrus; THA, thalamus.

Figure 3

Post hoc analyses of the strength of metabolic connectivity of the left triangle of the inferior frontal gyrus (IFGtri.L). Compared with the remaining three groups, metabolic connectivity between IFGtri.L and brain regions, including left ACC (ACC.L), left MCC (MCC.L), right MCC (MCC.R), left angular gyrus (ANG.L), left Heschl gyrus (HES.L), and right Heschl gyrus (HES.R) increased in the elderly patients with T2DM. AMYG, amygdala; CAL, calcarine; CAU, caudate; CUN, cuneus; FG, fusiform gyrus; HIP, hippocampus; INS, insula; IOG, inferior occipital gyrus; IPG, inferior parietal, but supramarginal and angular gyrus; ITG, inferior temporal gyrus; L, left; LING, lingual gyrus; med, medial; MFG, middle frontal gyrus; morb, medial orbital; MOG, middle occipital gyrus; MTG, middle temporal gyrus; OLF, olfactory cortex; oper, opercular part; orb, orbital part; p, temporal pole; PAL, pallidum; PCC, posterior cingulate cortex; PCL, paracentral lobule; PCUN, precuneus; PHIP, parahippocampal gyrus; PoCG, postcentral gyrus; PreCG, precental gyrus; PUT, putamen; R, right; REG, gyrus rectus; ROL, Rolandic operculum; SFG, superior frontal gyrus; SMA, supplementary motor area; SMG, supramarginal gyrus; SOG, superior occipital gyrus; SPG, superior parietal gyrus; STG, superior temporal gyrus; THA, thalamus.

Close modal
Table 2

Interaction effect between aging and T2DM on the strength of the metabolic connectivity of the left triangle of the inferior frontal gyrus

Seed ROIBrain regionSideFP
Left triangle of the inferior frontal gyrus ACC Left 4.949 0.028 
 MCC Left 7.326 0.008 
 MCC Right 4.753 0.031 
 Angular gyrus Left 5.572 0.020 
 Heschl gyrus Left 11.018 0.001 
 Heschl gyrus Right 4.222 0.042 
Seed ROIBrain regionSideFP
Left triangle of the inferior frontal gyrus ACC Left 4.949 0.028 
 MCC Left 7.326 0.008 
 MCC Right 4.753 0.031 
 Angular gyrus Left 5.572 0.020 
 Heschl gyrus Left 11.018 0.001 
 Heschl gyrus Right 4.222 0.042 

No Interaction Effect Between Aging and T2DM on Global Metabolic Network Properties

Neither main effect of aging nor interaction effect between aging and T2DM on the global metabolic network properties were found, including clustering coefficient (F = 1.557, P = 0.214), path length (F = 2.655, P = 0.106), global efficiency (F = 0.663, P = 0.417), local efficiency (F = 2.357, P = 0.127), σ (F = 0.215, P = 0.643), assortativity (F = 3.140, P = 0.079), synchronization (F = 1.154, P = 0.285), and hierarchy (F = 1.684, P = 0.197) (Table 3 and Supplementary Table 10). There were also no main effects of T2DM on these topological properties except assortativity (t = 4.452, P < 0.001), and patients with T2DM had higher global assortativity than HCs (mean difference 1.37 [95% CI 0.76, 1.97]) (Supplementary Table 11).

Table 3

Interaction effect between aging and T2DM on the global metabolic network properties

Topological propertyFP
Clustering coefficient 1.557 0.214 
Characteristic path length 2.655 0.106 
Global efficiency 0.663 0.417 
Local efficiency 2.357 0.127 
Sigma 0.215 0.643 
Assortativity 3.140 0.079 
Synchronization 1.154 0.285 
Hierarchy 1.684 0.197 
Topological propertyFP
Clustering coefficient 1.557 0.214 
Characteristic path length 2.655 0.106 
Global efficiency 0.663 0.417 
Local efficiency 2.357 0.127 
Sigma 0.215 0.643 
Assortativity 3.140 0.079 
Synchronization 1.154 0.285 
Hierarchy 1.684 0.197 

Increased Regional Metabolic Network Properties in Elderly Patients With T2DM

There were interaction effects between aging and T2DM on degree centrality of the left putamen (F = 4.600, P = 0.034), efficiency of the left ACC (F = 7.262, P = 0.008), and local efficiency of brain regions, including the right medial superior frontal gyrus (F = 4.225, P = 0.042), left calcarine (F = 4.799, P = 0.030), right precuneus (F = 3.946, P = 0.049), and left thalamus (F = 5.895, P = 0.017) (Table 4). The results of all post hoc analyses revealed that compared with elderly HCs and middle-aged patients with T2DM, regional metabolic network properties, except degree centrality, increased in the elderly patients with T2DM. On the contrary, degree centrality of the left putamen decreased in the elderly patients with T2DM (Supplementary Tables 12–17 and Fig. 4).

Figure 4

Post hoc analyses of regional metabolic network properties. Degree centrality of left putamen (A); efficiency of left ACC (B); and local efficiencies of the right medial superior frontal gyrus (C), left calcarine (D), right precuneus (E), and left thalamus (F). *P < 0.05; #P < 0.05.

Figure 4

Post hoc analyses of regional metabolic network properties. Degree centrality of left putamen (A); efficiency of left ACC (B); and local efficiencies of the right medial superior frontal gyrus (C), left calcarine (D), right precuneus (E), and left thalamus (F). *P < 0.05; #P < 0.05.

Close modal
Table 4

Interaction effect between aging and T2DM on the regional metabolic network properties

Topological propertyBrain regionSideFP
Degree centrality Putamen Left 4.600 0.034 
Efficiency ACC Left 7.262 0.008 
Local efficiency Superior frontal gyrus, medial Right 4.225 0.042 
 Calcarine Left 4.799 0.030 
 Precuneus Right 3.946 0.049 
 Thalamus Left 5.895 0.017 
Topological propertyBrain regionSideFP
Degree centrality Putamen Left 4.600 0.034 
Efficiency ACC Left 7.262 0.008 
Local efficiency Superior frontal gyrus, medial Right 4.225 0.042 
 Calcarine Left 4.799 0.030 
 Precuneus Right 3.946 0.049 
 Thalamus Left 5.895 0.017 

Partial Correlation Analyses

Partial correlation analysis showed that the glucose metabolism in the left triangle of the inferior frontal gyrus was negatively correlated with age (r = −0.198, P = 0.021) (Supplementary Fig. 2A) and FPG (r = −0.412, P < 0.001) (Supplementary Fig. 2B). The strength of metabolic connectivity between the left triangle of the inferior frontal gyrus and six brain regions (left ACC, bilateral MCCs, left angular gyrus, and bilateral Heschl gyri) had positive correlations with FPG (all P < 0.001) and age (all P < 0.05) (Supplementary Figs. 3 and 4). There was no significant partial correlation between glucose metabolism in the left triangle of the inferior frontal gyrus/strength of metabolic connectivity and BMI (all P > 0.05), and no significant correlation was found between regional metabolic network properties and age or FPG (all P > 0.05).

As a supplement to current knowledge, with a 2 × 2 factorial design, the current study is the first to investigate the interaction effect between aging and T2DM on brain glucose metabolism, metabolic connectivity, and most importantly, the patterns of the individual metabolic network rather than the group-level network. Considering potential confounding effects of other metabolic factors on the brain, the levels of BMI, blood pressure, and blood lipids in the four groups studied were also balanced (48,49). In the current study, we found that compared with the elderly HCs and middle-aged patients with T2DM, the elderly patients with T2DM showed decreased glucose metabolism in the triangle of the inferior frontal gyrus, and glucose metabolism in this brain region had a negative correlation with age and FPG. The JSD extends the asymmetrical KLSE to calculate a bounded and symmetrical divergence score of one probability distribution from another. The superiority of JSD over KLSE includes two aspects. First, the JSD score ranges from 0 (identical) to 1 (maximally different), resulting in a more accurate judgment of the similarity, while the KLSE score ranges from 0 to ∞. Second, JSD is symmetric, which makes it easier to portray connections between ROIs (40,50,51). On the basis of the individual metabolic network constructed with the novel JSD method, the current study for the first time reveals increased metabolic connectivity between the left triangle of the inferior frontal gyrus and several brain regions, including left ACC, bilateral MCCs, left angular gyrus, and bilateral Heschl gyri and increased efficiency in the ACC and local efficiency in the medial superior frontal gyrus, calcarine, precuneus, and thalamus in the elderly patients with T2DM. Metabolic connectivity changes induced by the interaction effect between aging and T2DM were positively correlated with age and FPG, which were opposite to the negative relationships between glucose metabolism in the triangle of the inferior frontal gyrus and age and FPG.

Decreased Neuronal Activities in the Frontal Lobe in Elderly Patients With T2DM

Both aging and T2DM have a long-term impact on the microvascular and macrovascular systems, which are associated with a high risk of neuropathy in the central nervous system (52,53). Consistent with previous studies, we found that compared with the elderly HCs and middle-aged patients with T2DM, the elderly patients with T2DM showed glucose hypometabolism in the triangle of the inferior frontal gyrus (54,55). The frontal lobe is an important brain region for high-level cognitive function and is particularly susceptible to vascular lesions (55,56). The imbalance between brain glucose uptake and oxygen utilization is aggravated with age and impairs the ability of brain aerobic glycolysis, which is related to the growth, formation, and remodeling of synapses (53,57). Ongoing low insulin levels in the brain might be an additional risk factor for cognitive impairment (58), as observed in previous studies in elderly patients with T2DM who had a relatively high risk of cognitive impairment, even dementia (8,59,60).

Strengthened Functional Integration and Increased Efficiency of Information Transmission Mainly Located in Cognition-Related Brain Regions in Elderly Patients With T2DM

In addition to studying localized changes in the brain, investigation of metabolic connectivity patterns with a multivariate nature can provide important supplementary information. Similar to the fMRI, PET is used to detect the functional status of neurons (61). Because of limitations of the network construction method, most previous studies constructed group-level metabolic covariance rather than individual metabolic connectivity and then compared the similarity or difference between metabolic covariance and functional connectivity (6264). However, the relationship between functional connectivity and individual metabolic connectivity has rarely been studied. Because PET depends less on neurovascular coupling, it can more directly reflect the neuronal function than fMRI. Therefore, metabolic connectivity might be an important complement to functional connectivity (65). The current study reveals that the brain metabolic connectivity of elderly patients with T2DM is characterized by bilateral changes. Evidence from other studies has demonstrated that brain regions showing a similar distribution of brain features have a coordinating relationship between them (66). Consistent with the concept of the JSD, the more similar the distribution of PDFs between any pair of brain regions, the smaller the divergence score and more synchronized the neuronal activities (17,40). Therefore, metabolic connectivity constructed with the JSD can be used as an edge measurement for functional integration between different brain regions. The increased strength of metabolic connectivity can be explained as strengthened functional integration. In the current study, we found strengthened functional integration between the triangle of the inferior frontal gyrus and several brain regions, including the left ACC, bilateral MCCs, left angular gyrus, and bilateral Heschl gyri, in the elderly patients with T2DM than that in the elderly HCs and middle-aged patients with T2DM. The underlying physiological basis for metabolic connectivity might be that the interconnected brain regions share a common distribution of major cell types (neurons, oligodendrocytes, and astrocytes) and gene expression related to synaptic transmission (67). However, the specific mechanism is unclear and needs further study.

The graphs offer a comprehensive map of how brain regions are connected, and the topological properties can be quantified with the graph theoretical analysis (43). The present results of the graph theoretical analysis show that compared with the elderly HCs and middle-aged patients with T2DM, the elderly patients with T2DM had increased regional metabolic network properties, manifested as increased efficiency and local efficiency in the ACC and several brain regions (medial superior frontal gyrus, calcarine, precuneus, and thalamus). According to the mathematical definitions, all topological properties, except for degree centrality, are calculated on the basis of the shortest path length (46). Thus, although topological properties were used to describe the architecture of metabolic networks from different aspects, the clinical implications that represent the efficiency of information communication of the brain at global and local levels have also been proven by other studies (6870). Therefore, on the basis of classical graph theory and previous neuroimaging studies, our results suggest that elderly patients with T2DM have a higher efficiency of information transmission mainly located in cognition-related brain regions, including the ACC, medial superior frontal gyrus, and precuneus, as well as the brain region responsible for visual function (calcarine), compared with elderly HCs and middle-aged patients with T2DM.

Strengthened Functional Integration and Increased Efficiency of Information Transmission Might Be Compensatory Changes for Brain Glucose Hypometabolism in Elderly Patients With T2DM

Insulin resistance is present in patients with T2DM and develops with age (2,71). A vicious cycle exists among insulin resistance, FPG, and glucagon level. Insulin resistance leads to an elevation of FPG levels, and increased FPG levels promote the secretion of glucagon, which in turn elevates the FPG levels, thus aggravating insulin resistance (72,73). In a previous study, insulin resistance was more severe in elderly participants with prediabetes or T2DM, and glucose metabolism was lower in the frontal lobe, parietotemporal lobe, and cingulate cortex (54). Consistently, we found in the current study that the glucose metabolism in the left triangle of the inferior frontal gyrus also decreased as aging and FPG increased. Conversely, in the individual metabolic network, the metabolic connectivity between the left triangle of the inferior frontal gyrus and six brain regions (left ACC, bilateral MCCs, left angular gyrus, and bilateral Heschl gyri) increased as aging and FPG increased. As interpreted in previous studies, strengthened functional integration and increased efficiency of information transmission mainly located in the cognition-related brain regions can also be explained as compensatory effects of brain glucose hypometabolism in elderly patients with T2DM.

In summary, we found that elderly patients with T2DM show glucose hypometabolism in the cognition-related brain region. From the insight of the individual metabolic network rather than the group-level network, the current study reveals for the first time that elderly patients with T2DM have strengthened functional integration and increased efficiency of information communication mainly located in the cognition-related brain regions and that these metabolic connectivity pattern changes might compensate for brain glucose hypometabolism. Because the neuroplastic property of the central nervous system can be an important basis to induce a behaviorally desirable clinical outcome, these findings might be potential therapeutic targets for plasticity-based therapy.

Y.-L.L. and J.-J.W. contributed equally to this work.

This article contains supplementary material online at https://doi.org/10.2337/figshare.19122653.

Acknowledgments. The authors thank Han Zhang from Shanghai Tech University for his assistance in the data processing for this article.

Funding. This work was supported by the National Key R&D Program of China (grants 2018YFC2001600 and 2018YFC2001604), National Natural Science Foundation of China (grants 81802249, 81871836, 81874035, and 81902301), Shanghai Science and Technology Committee (grants 18511108300, 18441903900, and 18441903800), Shanghai Rising-Star Program (grants 19QA1409000), Shanghai Municipal Commission of Health and Family Planning (grants 2018YQ02 and 201840224), Shanghai Youth Top Talent Development Plan and Shanghai “Rising Stars of Medical Talent” Youth Development Program (grant RY411.19.01.10), and Program of Shanghai Academic Research Leader (grant 19XD1403600).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.-L.L., J.-J.W., X.-Y.H., and M.-X.Z. contributed to the formal analysis. Y.-L.L. and M.-X.Z. wrote the original draft of the manuscript. J.-J.W., X.-Y.H., and M.-X.Z. contributed to the conception and methodology of the study. J.M., S.-S.L., X.X., D.W., and C.-L.S. contributed to the data validation. J.-G.X. reviewed and edited the manuscript. All authors read and approved the final manuscript. J.-G.X. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Tchkonia
T
,
Kirkland
JL
.
Aging, cell senescence, and chronic disease: emerging therapeutic strategies
.
JAMA
2018
;
320
:
1319
1320
2.
Chen
L
,
Magliano
DJ
,
Zimmet
PZ
.
The worldwide epidemiology of type 2 diabetes mellitus--present and future perspectives
.
Nat Rev Endocrinol
2011
;
8
:
228
236
3.
Meneilly
GS
,
Elliott
T
.
Metabolic alterations in middle-aged and elderly obese patients with type 2 diabetes
.
Diabetes Care
1999
;
22
:
112
118
4.
Huang
ES
.
Management of diabetes mellitus in older people with comorbidities
.
BMJ
2016
;
353
:
i2200
5.
Yang
W
,
Dall
TM
,
Beronjia
K
, et al.;
American Diabetes Association
.
Economic costs of diabetes in the U.S. in 2017
.
Diabetes Care
2018
;
41
:
917
928
6.
Sinclair
A
,
Saeedi
P
,
Kaundal
A
,
Karuranga
S
,
Malanda
B
,
Williams
R
.
Diabetes and global ageing among 65–99-year-old adults: findings from the International Diabetes Federation Diabetes Atlas, 9th edition
.
Diabetes Res Clin Pract
2020
;
162
:
108078
7.
Biessels
GJ
,
Despa
F
.
Cognitive decline and dementia in diabetes mellitus: mechanisms and clinical implications
.
Nat Rev Endocrinol
2018
;
14
:
591
604
8.
Roberts
RO
,
Knopman
DS
,
Geda
YE
, et al
.
Association of diabetes with amnestic and nonamnestic mild cognitive impairment
.
Alzheimers Dement
2014
;
10
:
18
26
9.
Chen
Y
,
Liu
Z
,
Zhang
J
, et al
.
Altered brain activation patterns under different working memory loads in patients with type 2 diabetes
.
Diabetes Care
2014
;
37
:
3157
3163
10.
Feinkohl
I
,
Keller
M
,
Robertson
CM
, et al
.
Cardiovascular risk factors and cognitive decline in older people with type 2 diabetes
.
Diabetologia
2015
;
58
:
1637
1645
11.
Kalyani
RR
,
Golden
SH
,
Cefalu
WT
.
Diabetes and aging: unique considerations and goals of care
.
Diabetes Care
2017
;
40
:
440
443
12.
van der Flier
WM
,
Skoog
I
,
Schneider
JA
, et al
.
Vascular cognitive impairment
.
Nat Rev Dis Primers
2018
;
4
:
18003
13.
Assar
ME
,
Angulo
J
,
Rodríguez-Mañas
L
.
Diabetes and ageing-induced vascular inflammation
.
J Physiol
2016
;
594
:
2125
2146
14.
Falvey
CM
,
Rosano
C
,
Simonsick
EM
, et al.;
Health ABC Study
.
Macro- and microstructural magnetic resonance imaging indices associated with diabetes among community-dwelling older adults
.
Diabetes Care
2013
;
36
:
677
682
15.
Biessels
GJ
,
Staekenborg
S
,
Brunner
E
,
Brayne
C
,
Scheltens
P
.
Risk of dementia in diabetes mellitus: a systematic review
.
Lancet Neurol
2006
;
5
:
64
74
16.
Hoge
RD
,
Pike
GB
.
Oxidative metabolism and the detection of neuronal activation via imaging
.
J Chem Neuroanat
2001
;
22
:
43
52
17.
Yakushev
I
,
Drzezga
A
,
Habeck
C
.
Metabolic connectivity: methods and applications
.
Curr Opin Neurol
2017
;
30
:
677
685
18.
Imai
M
,
Tanaka
M
,
Sakata
M
, et al
.
Metabolic network topology of Alzheimer’s disease and dementia with Lewy bodies generated using fluorodeoxyglucose positron emission tomography
.
J Alzheimers Dis
2020
;
73
:
197
207
19.
Arnemann
KL
,
Stöber
F
,
Narayan
S
,
Rabinovici
GD
,
Jagust
WJ
.
Metabolic brain networks in aging and preclinical Alzheimer’s disease
.
Neuroimage Clin
2017
;
17
:
987
999
20.
Wang
M
,
Jiang
J
,
Yan
Z
, et al.;
Alzheimer’s Disease Neuroimaging Initiative
.
Individual brain metabolic connectome indicator based on Kullback-Leibler divergence similarity estimation predicts progression from mild cognitive impairment to Alzheimer’s dementia
.
Eur J Nucl Med Mol Imaging
2020
;
47
:
2753
2764
21.
Wang
M
,
Yan
Z
,
Xiao
SY
,
Zuo
C
,
Jiang
J
.
A novel metabolic connectome method to predict progression to mild cognitive impairment
.
Behav Neurol
2020
;
2020
:
2825037
22.
Huang
SY
,
Hsu
JL
,
Lin
KJ
,
Hsiao
IT
.
A novel individual metabolic brain network for 18F-FDG PET imaging
.
Front Neurosci
2020
;
14
:
344
23.
Nielsen
F
.
On a generalization of the Jensen-Shannon divergence and the Jensen-Shannon centroid
.
Entropy (Basel)
2020
;
22
:
221
24.
Kong
XZ
,
Liu
Z
,
Huang
L
, et al
.
Mapping individual brain networks using statistical similarity in regional morphology from MRI
.
PLoS One
2015
;
10
:
e0141840
25.
Wang
H
,
Jin
X
,
Zhang
Y
,
Wang
J
.
Single-subject morphological brain networks: connectivity mapping, topological characterization and test-retest reliability
.
Brain Behav
2016
;
6
:
e00448
26.
Jiang
J
,
Zhou
H
,
Duan
H
, et al.;
Alzheimer’s Disease Neuroimaging Initiative
.
A novel individual-level morphological brain networks constructing method and its evaluation in PET and MR images
.
Heliyon
2017
;
3
:
e00475
27.
Malagurski
B
,
Liem
F
,
Oschwald
J
,
Mérillat
S
,
Jäncke
L
.
Longitudinal functional brain network reconfiguration in healthy aging
.
Hum Brain Mapp
2020
;
41
:
4829
4845
28.
van Bussel
FCG
,
Backes
WH
,
van Veenendaal
TM
, et al
.
Functional brain networks are altered in type 2 diabetes and prediabetes: signs for compensation of cognitive decrements? The Maastricht Study
.
Diabetes
2016
;
65
:
2404
2413
29.
Reijmer
YD
,
Leemans
A
,
Brundel
M
,
Kappelle
LJ
;
Utrecht Vascular Cognitive Impairment Study Group
.
Disruption of the cerebral white matter network is related to slowing of information processing speed in patients with type 2 diabetes
.
Diabetes
2013
;
62
:
2112
2115
30.
World Health Organization
.
Report of the World Health Organization. Active ageing: a policy framework
.
Aging Male
2002
;
5
:
1
37
31.
American Diabetes Association
.
2. Classification and diagnosis of diabetes: Standards of Medical Care in Diabetes-2019
.
Diabetes Care
2019
;
42
(
Suppl. 1
):
S13
S28
32.
Pi-Sunyer
FX
.
Obesity: criteria and classification
.
Proc Nutr Soc
2000
;
59
:
505
509
33.
Aklan
B
,
Oehmigen
M
,
Beiderwellen
K
, et al
.
Impact of point-spread function modeling on PET image quality in integrated PET/MR hybrid imaging
.
J Nucl Med
2016
;
57
:
78
84
34.
Liu
F
,
Jang
H
,
Kijowski
R
,
Bradshaw
T
,
McMillan
AB
.
Deep learning MR imaging-based attenuation correction for PET/MR imaging
.
Radiology
2018
;
286
:
676
684
35.
Van Aalst
J
,
Devrome
M
,
Van Weehaeghe
D
, et al
.
Regional glucose metabolic decreases with ageing are associated with microstructural white matter changes: a simultaneous PET/MR study
.
Eur J Nucl Med Mol Imaging
2022
;
49
:
664
680
36.
Hou
AL
,
Zheng
MX
,
Hua
XY
,
Huo
BB
,
Shen
J
,
Xu
JG
.
Electroacupuncture-related metabolic brain connectivity in neuropathic pain due to brachial plexus avulsion injury in rats
.
Front Neural Circuits
2020
;
14
:
35
37.
Huo
BB
,
Shen
J
,
Hua
XY
, et al
.
Alteration of metabolic connectivity in a rat model of deafferentation pain: a 18F-FDG PET/CT study
.
J Neurosurg
2019
;
132
:
1295
1303
38.
Tzourio-Mazoyer
N
,
Landeau
B
,
Papathanassiou
D
, et al
.
Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain
.
Neuroimage
2002
;
15
:
273
289
39.
Moon
YI
,
Rajagopalan
B
,
Lall
U
.
Estimation of mutual information using kernel density estimators
.
Phys Rev E Stat Phys Plasmas Fluids Relat Interdiscip Topics
1995
;
52
:
2318
2321
40.
Li
W
,
Tang
Y
,
Wang
Z
,
Hu
S
,
Gao
X
.
The reconfiguration pattern of individual brain metabolic connectome for Parkinson’s disease identification
.
29 April 2021 [preprint]. arXiv:2105.02811
41.
Wang
W-L
,
Li
Y-L
,
Zheng
M-X
, et al
.
Altered topological properties of grey matter structural covariance networks in complete thoracic spinal cord injury patients: a graph theoretical network analysis
.
Neural Plast
2021
;
2021
:
8815144
42.
Wang
J
,
Wang
X
,
Xia
M
,
Liao
X
,
Evans
A
,
He
Y
.
GRETNA: a graph theoretical network analysis toolbox for imaging connectomics
.
Front Hum Neurosci
2015
;
9
:
386
43.
Rubinov
M
,
Sporns
O
.
Complex network measures of brain connectivity: uses and interpretations
.
Neuroimage
2010
;
52
:
1059
1069
44.
He
Y
,
Dagher
A
,
Chen
Z
, et al
.
Impaired small-world efficiency in structural cortical networks in multiple sclerosis associated with white matter lesion load
.
Brain
2009
;
132
:
3366
3379
45.
Sporns
O
.
The human connectome: a complex network
.
Ann N Y Acad Sci
2011
;
1224
:
109
125
46.
Bullmore
ET
,
Bassett
DS
.
Brain graphs: graphical models of the human brain connectome
.
Annu Rev Clin Psychol
2011
;
7
:
113
140
47.
Krzywinski
M
,
Schein
J
,
Birol
I
, et al
.
Circos: an information aesthetic for comparative genomics
.
Genome Res
2009
;
19
:
1639
1645
48.
Eckel
RH
,
Grundy
SM
,
Zimmet
PZ
.
The metabolic syndrome
.
Lancet
2005
;
365
:
1415
1428
49.
Alberti
KGMM
,
Eckel
RH
,
Grundy
SM
, et al.;
International Diabetes Federation Task Force on Epidemiology and Prevention
;
National Heart, Lung, and Blood Institute
;
American Heart Association
;
World Heart Federation
;
International Atherosclerosis Society
;
International Association for the Study of Obesity
.
Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity
.
Circulation
2009
;
120
:
1640
1645
50.
Murphy
KP
.
Machine Learning: A Probabilistic Perspective
.
Cambridge, MA
,
The MIT Press
,
2012
, pp.
56
61
51.
Lin
J
.
Divergence measures based on the Shannon entropy
.
IEEE Trans Inf Theory
1991
;
37
:
145
151
52.
Brundel
M
,
Kappelle
LJ
,
Biessels
GJ
.
Brain imaging in type 2 diabetes
.
Eur Neuropsychopharmacol
2014
;
24
:
1967
1981
53.
Goyal
MS
,
Vlassenko
AG
,
Blazey
TM
, et al
.
Loss of brain aerobic glycolysis in normal human aging
.
Cell Metab
2017
;
26
:
353
360.e3
54.
Baker
LD
,
Cross
DJ
,
Minoshima
S
,
Belongia
D
,
Watson
GS
,
Craft
S
.
Insulin resistance and Alzheimer-like reductions in regional cerebral glucose metabolism for cognitively normal adults with prediabetes or early type 2 diabetes
.
Arch Neurol
2011
;
68
:
51
57
55.
Kuczynski
B
,
Jagust
W
,
Chui
HC
,
Reed
B
.
An inverse association of cardiovascular risk and frontal lobe glucose metabolism
.
Neurology
2009
;
72
:
738
743
56.
Fellows
LK
.
The functions of the frontal lobes: evidence from patients with focal brain damage
.
Handb Clin Neurol
2019
;
163
:
19
34
57.
Goyal
MS
,
Hawrylycz
M
,
Miller
JA
,
Snyder
AZ
,
Raichle
ME
.
Aerobic glycolysis in the human brain is associated with development and neotenous gene expression
.
Cell Metab
2014
;
19
:
49
57
58.
Eriksson
O
,
Långström
B
,
Antoni
G
.
News ways of understanding the complex biology of diabetes using PET
.
Nucl Med Biol
2021
;
92
:
65
71
59.
Crane
PK
,
Walker
R
,
Hubbard
RA
, et al
.
Glucose levels and risk of dementia
.
N Engl J Med
2013
;
369
:
540
548
60.
Xue
M
,
Xu
W
,
Ou
YN
, et al
.
Diabetes mellitus and risks of cognitive impairment and dementia: A systematic review and meta-analysis of 144 prospective studies
.
Ageing Res Rev
2019
;
55
:
100944
61.
Magistretti
PJ
.
Neuron-glia metabolic coupling and plasticity
.
J Exp Biol
2006
;
209
:
2304
2311
62.
Passow
S
,
Specht
K
,
Adamsen
TC
, et al
.
Default-mode network functional connectivity is closely related to metabolic activity
.
Hum Brain Mapp
2015
;
36
:
2027
2038
63.
Savio
A
,
Fünger
S
,
Tahmasian
M
, et al
.
Resting-state networks as simultaneously measured with functional MRI and PET
.
J Nucl Med
2017
;
58
:
1314
1317
64.
Di
X
;
Alzheimer’s Disease Neuroimaging Initiative
.
Metabolic brain covariant networks as revealed by FDG-PET with reference to resting-state fMRI networks
.
Brain Connect
2012
;
2
:
275
283
65.
Jamadar
SD
,
Ward
PGD
,
Liang
EX
,
Orchard
ER
,
Chen
Z
,
Egan
GF
.
Metabolic and hemodynamic resting-state connectivity of the human brain: a high-temporal resolution simultaneous BOLD-fMRI and FDG-fPET multimodality study
.
Cereb Cortex
2021
;
31
:
2855
2867
66.
Alexander-Bloch
A
,
Raznahan
A
,
Bullmore
E
,
Giedd
J
.
The convergence of maturational change and structural covariance in human cortical networks
.
J Neurosci
2013
;
33
:
2889
2899
67.
Hawrylycz
MJ
,
Lein
ES
,
Guillozet-Bongaarts
AL
, et al
.
An anatomically comprehensive atlas of the adult human brain transcriptome
.
Nature
2012
;
489
:
391
399
68.
Vecchio
F
,
Miraglia
F
,
Bramanti
P
,
Rossini
PM
.
Human brain networks in physiological aging: a graph theoretical analysis of cortical connectivity from EEG data
.
J Alzheimers Dis
2014
;
41
:
1239
1249
69.
Vecchio
F
,
Miraglia
F
,
Maria Rossini
P
.
Connectome: graph theory application in functional brain network architecture
.
Clin Neurophysiol Pract
2017
;
2
:
206
213
70.
Qiu
X
,
Zhang
Y
,
Feng
H
,
Jiang
D
.
Positron emission tomography reveals abnormal topological organization in functional brain network in diabetic patients
.
Front Neurosci
2016
;
10
:
235
71.
He
M
,
Chiang
HH
,
Luo
H
, et al
.
An acetylation switch of the NLRP3 inflammasome regulates aging-associated chronic inflammation and insulin resistance
.
Cell Metab
2020
;
31
:
580
591.e5
72.
Færch
K
,
Vistisen
D
,
Pacini
G
, et al
.
Insulin resistance is accompanied by increased fasting glucagon and delayed glucagon suppression in individuals with normal and impaired glucose regulation
.
Diabetes
2016
;
65
:
3473
3481
73.
Saltiel
AR
,
Kahn
CR
.
Insulin signalling and the regulation of glucose and lipid metabolism
.
Nature
2001
;
414
:
799
806
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.