Abnormal spontaneous brain activity in type 2 diabetes with and without microangiopathy revealed by regional homogeneity

European Journal of Radiology 85 (2016) 607–615

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

Abnormal spontaneous brain activity in type 2 diabetes with and without microangiopathy revealed by regional homogeneity Juan Peng a , Hang Qu a , Jing Peng b , Tian-You Luo a,∗ , Fa-Jin Lv a , Li chen a , Zhuo-Nan Wang a , Yu Ouyang a , Qing-Feng Cheng c a

Department of Radiology, the First Affiliated Hospital, Chongqing Medical University, No. 1 Youyi Road, Yuzhong District, Chongqing 400016, China Department of Clinical Laboratory, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030 China c Department of Endocrinology, the First Affiliated Hospital, Chongqing Medical University, No. 1 Youyi Road, Yuzhong District, Chongqing 400016, China b

a r t i c l e

i n f o

Article history: Received 18 August 2015 Received in revised form 12 December 2015 Accepted 27 December 2015 Keywords: Type 2 diabetes Cognitive impairment Regional homogeneity Resting-state fMRI Microangiopathy

a b s t r a c t Purpose: To investigate whether global spontaneous brain activity changes in type 2 diabetes mellitus (T2DM) patients and these changes vary according to the degree of microangiopathy. Materials and methods: T2DM patients with (M+ , n = 26) and without (M− , n = 22) microangiopathy as well as 28 healthy nondiabetic subjects were enrolled in this study. All the subjects completed a restingstate functional magnetic resonance imaging (rs-fMRI) examination and neuropsychological assessment. Regional homogeneity (ReHo) values, representing spontaneous brain activity, were calculated and compared between M+ and M− T2DM patients and nondiabetic controls. Results: In both M+ and M− T2DM patients, ReHo values were decreased in the occipital lobe, temporal lobe, postcentral gyrus, and cerebellum, while increased in the bilateral precuneus, superior/middle frontal gyrus, and insula. Compared with the M− group, M+ patients showed decreased ReHo values in the left cuneus and superior occipital gyrus. The ReHo values in the lingual gyrus/calcarine cortex and MTG were related to clinical parameters in T2DM patients. Conclusion: The abnormalities of spontaneous brain activity revealed by ReHo values in both M+ and M− T2DM patients may provide insights into the neurological pathophysiology underlying diabetes-related cognitive impairments. M+ patients showed more decreased brain activity related to severely impaired function of visual processing and visual memory. © 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Type 2 diabetes mellitus (T2DM) is associated with impaired cognition which mainly presented as declined mental speed, attention, memory, and executive function [1], although the exact pathophysiological mechanism of T2DM-induced cognitive

Abbreviations: FG, fusiform gyrus; IPL, inferior parietal lobe; MFG, middle frontal gyrus; MOG, middle occipital gyrus; MPFC, medial prefrontal cortex; MTG, middle temporal gyrus; PCC, posterior cingulate cortex; PCu, precuneus; PoCG, post-central gyrus; PrCG, pre-central gyrus; SFG, superior frontal gyrus; SOG, superior occipital gyrus; STG, superior temporal gyrus. ∗ Corresponding author. Fax: +86 23 68811487. E-mail addresses: [email protected] (J. Peng), [email protected] (H. Qu), [email protected] (J. Peng), [email protected] (T.-Y. Luo), [email protected] (F.-J. Lv), [email protected] (L. chen), [email protected] (Z.-N. Wang), [email protected] (Y. Ouyang), [email protected] (Q.-F. Cheng). http://dx.doi.org/10.1016/j.ejrad.2015.12.024 0720-048X/© 2015 Elsevier Ireland Ltd. All rights reserved.

impairment has not yet been fully elucidated. Longitudinal epidemiological studies have shown that chronic hyperglycaemia and microangiopathy are associated with increased risk of cognitive dysfunction in both type 2 and type 1 diabetes [2,3]. Since the retinal and cerebral microvasculature share some common features, including similar embryological origin, size, and structure, and physiological characteristics, microangiopathy in the retina might be an indirect marker for changes in the cerebral microangiopathy [4]. Resting-state functional magnetic resonance imaging (rs-fMRI) has become a promising technique to exhibit neurophysiological mechanism in various neuropsychiatric disorders [5–8]. Recently, rs-fMRI has been used to investigate whether altered spontaneous brain activity exists in DM patients. Musen et al. observed that T2DM patients had abnormal functional connectivity among several brain regions in the default mode network (DMN) when compared with controls [9]. Meanwhile, Xia et al. demonstrated altered amplitude of low-frequency fluctuations (ALFF) in some

608

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615

brain areas in T2DM patients which was correlated with declined neurocognitive function, severity of hyperglycemic state and damaged ␤-cell function [10]. In a previous rs-fMRI study, van Duinkerken et al. suggested that cognitive decrements in T1DM patients were related to changes in resting-state neural connectivity and that these alterations dependent on the degree of microangiopathy [11]. Regional homogeneity (ReHo) analysis is an important method for depicting the various characteristics of global rs-fMRI signals, in which Kendall coefficient of concordace (KCC) was used to measure the similarity of the time series of a given voxel to those of its nearest neighbors in a voxel-wise way [12]. It has been widely used to evaluate the functional abnormities in various diseases such as Parkinson’s disease [6], major depression [7], Alzheimer’s disease [8], and essential tremor [13]. However, few studies [14] have been conducted to investigate whether global spontaneous brain activity changes in T2DM patients and these changes vary according to the degree of microangiopathy using ReHo techniques. In this study, ReHo analysis was used to assess whether spontaneous brain activity changed prior to structural abnormalities in T2DM patients. We hypothesized that (1) cognitive decrements in T2DM patients are associated with alterations in ReHo values and that these changes may vary according to the degree of microangiopathy; and (2) the spontaneous brain activity changes would be correlated with cognitive dysfunction and T2DM-related clinical parameters.

2.2. Clinical data and cognitive assessment Clinical data were recorded, including blood pressure, weight and height, body mass index (BMI) = (weight in kg)/(height in m)2 . Blood samples were obtained by venepuncture at 8 A.M. after overnight fasting to assess the levels of fasting blood glucose (FBG), glycosylated hemoglobin (HbAlc), total cholesterol, triglyceride and low density lipoprotein cholesterol. A battery of neuropsychological assessments was completed, which based on previous studies concerning cognitive impairment in T2DM patients [10,19,20]. The Mini Mental State Examination (MMSE), which usually used to screen for dementia, was included as a general psychological assessment [19]. The Trail-Making Test (TMT) Parts A and B (TMT-A and -B) were used to evaluate attention, psychomotor speed, and executive function; these parameters were then assessed with TMT, which is calculated as the difference between the times for each part (TMT-B minus TMT-A) and thought to be a more accurate measure of executive functions than performance on TMT-A or TMT-B [20]. The Auditory Verbal Learning Test (AVLT), a test of verbal memory, assesses registration and recall of words, while Rey–Osterrieth Complex Figure Test (CFT) assesses visuospatial memory and visual spatial skills [10,14]. The Clock Drawing Test (CDT) was used to evaluate executive function and visual spatial skills [10]. All of the tests took approximately 50 min to complete. 2.3. MRI data acquisition

2. Materials and methods 2.1. Subjects The present study was conducted from February 2012 to October 2013 and was approved by the institutional review board. Written informed consents were obtained from all the subjects. Eighty-five participants were enrolled in this study, including 29 T2DM patients with microangiopathy (M+ ) and 26 patients without microangiopathy (M− ) and 30 nondiabetic subjects. Three of the M+ patients, four of the M− patients and two of the nondiabetic controls were excluded because their head motion was out of range during the MRI data acquisition. The patients were recruited from the endocrinology department of the affiliated hospital of our university, while the control subjects were from the local community during the same period. All the participants were at least 40 years of age, receiving at least 5 years of education and right handed. T2DM was diagnosed according to the latest criteria published by the American Diabetes Association [15]. M+ patients were chosen on the basis of diabetic retinopathy, but other microangiopathy, such as microalbuminuria and peripheral neuropathy, could also be accompanied, while M− patients were free of any of the microvascular complications. Diabetic retinopathy was diagnosed using fundus photography, which was assessed according to the EURODIAB classification criteria [16] and patients with scores from 1 (minimal non-proliferative retinopathy) to 5 (proliferative retinopathy) were included in the M+ group. Microalbuminuria was defined by an albumin/creatinine ratio (ACR) >30 mg/gCr [17]. Peripheral neuropathy was assessed for all subjects in accordance with the Diabetes Control and Complications Trial criteria using electrophysiology tests [18]. Exclusion criteria for all participants included a history of T2DMrelated acute metabolic complications, severe hypoglycemia episode, stroke, epilepsy, dementia, head injury, major depression or other psychiatric illness, major medical illness (e.g., anemia, cancer and thyroid dysfunction), alcoholism or drug abuse, severe visual or hearing loss, and contraindication for MRI.

MRI scanning was performed on a GE Signa Hdxt 3.0 T scanner (General Electric Medical Systems, USA) using an eight-channel phased-array head coil. Foam padding was used to restrict head movement and ear plugs were used to minimise scanner noise. During rs-fMRI acquisition, all subjects were instructed to keep their eyes closed and not to think of anything. Rs-fMRI data were acquired using an echo-planar image (EPI) pulse sequence with parameters as follows: TR = 2000 ms, TE = 40 ms, flip angle = 90◦ , thickness/gap = 4.0/0 mm, FOV = 240 × 240 mm, and matrix = 64 × 64. A total of 240 time points were obtained in 8 min. High-resolution 3D-T1-weighted axial image (TR = 8.3 ms, TE = 3.3 ms, flip angle = 15◦ , thickness/gap = 1.0/0 mm, FOV = 240 × 240 mm, and matrix = 256 × 192) and T2-FLAIR-weighted image (TR = 8000 ms, TE = 126 ms, TI = 1500 ms, thickness/gap = 5.0/1.5 mm, FOV = 240 × 240 mm, and matrix = 256 × 192) were also acquired. 2.4. VBM analysis To check out whether there were structural differences between M+ and M− T2DM patients and control subjects, the high-resolution 3D-T1-weighted images were processed using the VBM8 toolbox software (http://dbm.neuro.uni-jena.de/vbm). In brief, the T1weighted images were segmented into gray matter (GM), whiter matter (WM), and cerebrospinal fluid (CSF) using the unified segmentation model. After that, one-way analysis of variance (ANOVA) with false discovery rate (FDR) corrections was carried out to evaluate between-group differences in GM and WM volume. 2.5. Data preprocessing and ReHo analysis We excluded participants with obvious lacunar infarction, cerebral atrophy or WM lesions, which were assessed separately by two experienced radiologists on conventional MRI. After careful evaluation of the quality of raw functional images, data preprocessing, including slice timing, realignment, and normalization, was performed with the Data Processing Assistant for rs-fMRI (DPARSF; http://www.restfmri.net/forum/DPARSF) through SPM8 software

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615

609

Fig. 1. Representative one-sample t test results of ReHo values in M+ (A), M− (B)T2DM patients, and nondiabetic controls (C) (P < 0.05, AlphaSim corrected). R, right; L, left. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

(http://www.fil.ion.ucl.ac.uk/spm/) and REST software (REST1.8; http://www.restfmri.net). The first five volumes (10 s) of the data were discarded to ensure stable magnetization and allow the participants to adapt to the EPI scanning environment. Any subjects with head motion exceeding 3.0-mm for translation or 3.0◦ for rotation in any direction were excluded. The functional images were then spatially normalized to standard coordinates and resampled to 3 × 3 × 3 mm3 . After that, the linear trend of the time series was removed and a temporal

filter (0.01 Hz < f < 0.08 Hz) was conducted to reduce the effects of low-frequency drift and physiological high-frequency noise. We used Kendall’s coefficient of concordance (KCC, also called ReHo value) [21] to measure local homogeneity or similarity of the ranked time series of a given voxel to its nearest 26 neighbor voxels in voxel wise manner [12]:

 W=

(Ri )2 − n(R)

1 2 3 K (n 12

2

− n)

Table 1 Demographic, clinical, and cognitive characteristics of participants.

N Age (years) Sex (male/female)a Education (years) Disease duration (years) BMI (kg/m2 ) Fasting glucose (mmol/l) HbA1 c (% [mmol/mol]) Systolic BP (mmHg) Diastolic BP (mmHg) Hypertension (%)a Dyslipidemia (%)a ACR (mg/gCr) Microalbuminuria (%) Neuropathy (%) Mini-Mental State Exam Trail Making Test-A (s) Trail Making Test-B (s) TMT (s) AVLT Clock Drawing Test CFT-delayed recall (20 min)

M+ T2DM

M− T2DM

Control

P

26 57.6 ± 9.3 12/14 10.3 ± 2.9 12.1 ± 5.8 24.3 ± 3.5 9.5 ± 4.0 8.8 ± 1.6 (72.2 ± 19.6) 134.8 ± 12.6 80.6 ± 9.4 10 (38.5) 9 (34.6) 558.8 ± 187.2 20 (76.9) 23 (88.4) 28.8 ± 0.6 73.2 ± 12.7 183.1 ± 31.4 109.9 ± 33.8 31.3 ± 3.2 3.5 ± 0.5 13.8 ± 2.7

22 58.8 ± 7.9 10/12 10.0 ± 2.1 10.9 ± 3.4 23.8 ± 3.2 8.6 ± 2.8 8.1 ± 2.2 (65.0 ± 23.5) 132.7 ± 7.4 82.3 ± 7.0 8 (36.4) 8 (36.4) 12.6 ± 10.4

28 56.2 ± 6.9 12/16 10.4 ± 2.1 – 24.1 ± 3.3 5.2 ± 0.35 5.5 ± 0.3 (35.6 ± 3.5) 131.9 ± 6.4 81.3 ± 5.4 8 (28.6) 9 (32.1) 12.0 ± 4.3

– 0.556 0.968 0.516 0.382 0.675 <0.001* <0.001*

28.9 ± 0.7 71.2 ± 13.1 177.8 ± 23.8 106.6 ± 27.1 31.5 ± 5.9 3.6 ± 0.5 14.4 ± 2.2

29.2 ± 0.6 62.1 ± 12.3 149.3 ± 25.1 87.1 ± 27.9 32.6 ± 1.8 3.6 ± 0.7 19.1 ± 1.6

0.135 0.001* 0.001* 0.001* 0.132 0.536 0.001*

Data are mean ± SD or absolute numbers (%). AVLT, Auditory Verbal Learning Test. TMT, TMT-B minus TMT-A. a The P value for proportions was obtained by x2 test. ACR, albumin/creatinine ratio. Microalbuminuria was defined as an ACR ≥30 mg/gCr. * P < 0.05.

0.551 0.753 0.723 0.951 0.001*

610

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615

where W is the KCC among given voxels, ranging from 0 to 1; Ri is the sum rank of the ith time point; R = (n + 1) K/2 is the mean of the Ri s; K is the number of time series within a measured cluster (27, one given voxel plus the number of its neighbors); n is the number of ranks (here, n = 240 time points). The individual ReHo map was generated in a voxel-wise way using the REST software. After that, a mask was made from the MNI template. To achieve standardization, each individual ReHo map was divided by its own mean ReHo within the mask [12]. To reduce noise and residual differences in gyral anatomy, the ReHo maps were smoothed with a 4-mm full-width at half-maximum (FWHM) Gaussian kernel. 2.6. Statistical analysis 2.6.1. Demographic and clinical characteristics Demographic and clinical parameters and cognitive test scores were compared among the three groups using SPSS software (version 18.0, SPSS Inc., Chicago, IL). One-way ANOVA was used to assess differences in continuous variables across the groups and a x2 test was used for proportions. P values < 0.05 were considered statistically significant. 2.6.2. Within-group ReHo analysis To explore the within-group ReHo patterns, one-sample t tests were performed on the individual ReHo maps in a voxel-wise way for each group. The within-group statistical threshold was set at P < 0.01 and cluster size >40 voxels, which corresponded to a corrected P < 0.05, corrected with Alphasim program criterion (http:// afni.nih.gov/afni/docpdf/AlphaSim.pdf). 2.6.3. Between-group ReHo analysis To investigate the ReHo differences among the three groups, one-way ANOVA was performed at each voxel. If statistical difference was present, post hoc individual tests were further conducted to detect differences between each pair of the three groups on individual ReHo maps. Age, gender, education, BMI, hyperlipidemia, and hypertension were imported as covariates in the ANOVA and following post-hoc statistic analysis to avoid any confounding effects. To eliminate the effect of structural differences on ReHo analysis, the resulting GM volume maps from VBM analysis were also entered as covariates in the functional data analysis. The statistical threshold for the group analysis was set at P < 0.01 at the voxel level with a minimum cluster size of 40 voxels, which corresponded to a corrected P < 0.05 (AlphaSim corrected). 2.6.4. Clinical correlation analysis To investigate which brain regions correlated with the clinical parameters, voxel-wise correlation analysis was performed between ReHo values at each voxel in the whole brain and the clinical, biochemical, or neuropsychological parameters with the REST software, adjusted for age, gender, education, BMI, hyperlipidemia, and hypertension. The threshold was set at P < 0.05 (AlphaSim corrected) with P < 0.01 at the voxel level and a minimum cluster size of 40 voxels. 3. Results 3.1. Demographic and cognitive characteristics Demographics, clinical and neuropsychological data for T2DM patients and nondiabetic controls are summarized in Table 1. Both M+ and M− patients had increased FPG and HbAlc levels compared with control groups, however, no differences of these measurements were observed between M+ and M− patients. Compared with M− patients and nondiabetic controls, M+ patients had an increased

Table 2 Regions showing ReHo differences between M+ and M− T2DM patients and nondiabetic controls. Brain regions

B lingual gyrus B calcarine cortex B SOG/MOG/FG B cuneus/precuneus B insula B SFG/MFG L STG/MTG R STG/MTG L postcentral gyrus R postcentral gyrus B cerebellum

BA

17/18 17/18 18/19 7 8/9 42/37 21/48 3 3

MNI coordinates(mm)

Peak F value

Voxels

−3

31.880

9417

9 −27 33 39 −57

30.651 27.017 16.909 17.320 17.341

680 1666 488 205 534

x

y

z

0

−75

−54 45 −45 42 −21

−24 −3 −18 −18 −63

BA, Brodmann area; B: bilateral; L, left; R, right; MNI, Montreal Neurological Institute; x, y, z, coordinates of primary peak locations in the MNI space; SOG, superior occipital gyrus; MOG, middle occipital gyrus; FG, fusiform gyrus; SFG, superior frontal gyrus; MFG, middle frontal gyrus; STG, superior temporal gyrus; MTG, middle temporal gyrus; P < 0.05, AlphaSim corrected.

ACR, twenty of these patients had microalbuminuria, and twentythree had neuropathy. There were no significant differences in age, sex, education, disease duration, BMI, blood pressure, and blood lipids among the three groups. In terms of cognitive assessment, both M+ and M− patients performed more poorly than control groups on the TMT-A, TMTB, TMT and CFT-delayed recall tests. Although no significant differences in cognitive performance as determined by the aforementioned methods were found between M+ and M− patients, the M+ group showed slight decreases in cognitive performance on the above tests compared with the M− group. No significant differences were observed in the MMSE, AVLT and CDT scores among the three groups. 3.2. VBM analysis Although the GM and WM volumes of M+ and M− patients decreased slightly, they did not significantly differ from those of the healthy controls (one-way ANOVA, no survived clusters, FDR corrected). 3.3. ReHo analysis One-sample t tests were used to compare the ReHo values within each group. In the three groups, ReHo values in the posterior cingulate cortex (PCC)/precuneus (PCu), medial prefrontal cortex (MPFC), and bilateral inferior parietal lobe (IPL) were significantly higher than the global mean values (P < 0.05, AlphaSim corrected). The ReHo patterns were in line with those of the default mode network (DMN) revealed in previous studies [22]. Additionally, similar changes could also be observed in several other regions, such as the temporal lobes, occipital regions, pre- and post-central gyrus (PrCG and PoCG) (Fig. 1). For determining ReHo differences among the three groups, one-way ANOVA was used. Significant between-group differences in ReHo values were observed (P < 0.05, AlphaSim corrected) in the occipital regions (i.e., bilateral lingual gyrus/calcarine cortex, superior occipital gyrus (SOG), middle occipital gyrus (MOG), and cuneus) and temporal regions such as bilateral superior temporal gyrus (STG), middle temporal gyrus (MTG), and fusiform gyrus (FG), as well as the parietal regions such as the PCu and PoCG. In addition, differences were also found in the frontal regions such as the superior frontal gyrus (SFG), middle frontal gyrus (MFG), insula, and cerebellum (Fig. 2A, Table 2).

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615

611

Fig. 2. Analytical results for the ReHo differences between M+ and M− T2DM patients and nondiabetic controls (P < 0.05, AlphaSim corrected). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) (A) ReHo differences between the three groups based on one-way ANOVA analysis. (B) ReHo differences between M+ patients and nondiabetic controls. (C) ReHo differences between M− patients and nondiabetic controls. (D) ReHo differences between M+ and M− patients. Red and blue regions indicate those with increased and decreased ReHo values, respectively, when the former groups were compared to the latter ones. R, right; L, left.

After the one-way ANOVA analysis, the post hoc individual tests were used to compare the ReHo values between each pair of the three groups. Compared with the nondiabetic controls, both M+ and M− patients showed decreased ReHo values in the occipital lobe (bilateral lingual gyrus/calcarine cortex, left SOG, MOG, and cuneus), temporal lobe (bilateral STG, MTG, and FG), parietal lobe (bilateral PoCG), and bilateral cerebellum. However, ReHo values in the bilateral PCu, SFG, MFG, and insula were significantly higher in M+ and M− patients than in nondiabetic controls (P < 0.05, AlphaSim corrected) (Fig. 2B, Table 3 and Fig. 2C, Table 4 respectively). Compared with the M− group, M+ patients showed decreased ReHo values in the left cuneus and SOG (P < 0.05, AlphaSim corrected) (Fig. 2D, Table 5). 3.4. Correlation analysis Voxel-wise correlation analysis revealed that ReHo values in the lingual gyrus/calcarine cortex were negatively related to BMI, TMTB and TMT, while positively related to CFT-delayed score in all T2DM patients. In addition, ReHo values in the MTG were negatively correlated with BMI, TMT-B and TMT (all P < 0.05, AlphaSim corrected; Fig. 3, Supplementary Table 1). However, no significant

Table 3 Regions showing ReHo differences between M+ T2DM patients and nondiabetic controls. Brain regions

Decreased in M+ B lingual gyrus B calcarine cortex L SOG/MOG/cuneus B fusiform gyrus L STG/MTG R STG B postcentral gyrus B cerebellum Increased in M+ B SFG/MFG B precuneus B insula

BA

17/18 17/18 18/19 19 42/37 48 3

8/9 7

MNI coordinates(mm)

Peak tvalue

Voxels

0

−7.139

1632

−31 −18

15 6

−4.745 −7.260

200 212

0

−63

−51

−5.390

400

−36 21

27 −63

54 33

6.979 7.291

1082 1923

x

y

z

12

−54

−61 48

BA, Brodmann area; B: bilateral; L, left; R, right; MNI, Montreal Neurological Institute; x, y, z, coordinates of primary peak locations in the MNI space; SOG, superior occipital gyrus; MOG, middle occipital gyrus; STG, superior temporal gyrus; MTG, middle temporal gyrus; SFG, superior frontal gyrus; MFG, middle frontal gyrus; P < 0.05, AlphaSim corrected.

612

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615

Fig. 3. Representative results of correlation analysis based on a voxel-wise analysis in T2DM patients (P < 0.05, AlphaSim corrected). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) ReHo values in the lingual gyrus/calcarine cortex and MTG were negatively related to BMI (A), TMT-B (B) and TMT (C). Significant positive correlations were found between the ReHo values in lingual gyrus/calcarine cortex and CFT-delayed scores (D). R, right. Hot and cold colors indicate positive and negative correlations, respectively, between ReHo values and clinical parameters.

correlations were found between any regional ReHo values and age, education level, HbA1c, disease duration, and other factors. We also performed correlation analysis between ReHo values and the clinical features in the control group, which showed no significant correlation between any regional ReHo values and clinical factors.

4. Discussion The present study reveals that the ReHo approach can detect patterns of spontaneous brain activity changes in T2DM patients

in the resting state. In both M+ and M− patients, ReHo values were decreased in the occipital lobe, temporal lobe, PoCG, and cerebellum, while increased in bilateral PCu, SFG, MFG and insula compared to the nondiabetic controls, suggesting that neural function in certain areas is less or more synchronized in T2DM patients. Compared with the M− group, M+ patients showed decreased ReHo values in left cuneus and SOG. The ReHo values in the lingual gyrus/calcarine cortex and MTG were related to clinical variables in T2DM patients.

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615 Table 4 Regions showing ReHo differences between M− T2DM patients and nondiabetic controls. Brain regions

Decreased in M− B lingual gyrus B calcarine cortex L IOG/cuneus B fusiform gyrus L STG R STG B postcentral gyrus B cerebellum

BA

MNI coordinates(mm) x

y

z

Peak tvalue

Voxels

17/18 17/18 18/19 19 48 48 3/2

0

−75

−3

−7.139

1213

−54 45 −34

−24 −19 −28

9 13 53

−6.813 −5.516 −4.570

223 212 140

8/9 7

−17 12

37 −42

57 9

5.623 6.481

670 1708

Increased in M− B SFG/MFG B precuneus B insula

BA, Brodmann area; B: bilateral; L, left; R, right; MNI, Montreal Neurological Institute; x, y, z, coordinates of primary peak locations in the MNI space; IOG, inferior occipital gyrus; STG, superior temporal gyrus; SFG, superior frontal gyrus; MFG, middle frontal gyrus; P < 0.05, AlphaSim corrected.

Table 5 Regions showing ReHo differences between M+ and M− T2DM patients. Brain regions

Decreased in M+ L cuneus/SOG

BA

7

MNI coordinates(mm) x

y

z

−15

−75

36

Peak tvalue

Voxels

−4.673

91

BA, Brodmann area; L, left; MNI, Montreal Neurological Institute; x, y, z, coordinates of primary peak locations in the MNI space; SOG, superior occipital gyrus; P < 0.05, AlphaSim corrected.

Compared with the nondiabetic controls, both M+ and M− patients showed decreased ReHo values in the occipital lobe, which correlated with the original stage of visual processing [23]. A previous fMRI study by Wang et al. showed that the neural network associated with the spontaneous activity of the primary visual area (PVA), including the visual association areas, PrCG, PoCG, MFG, FG, MTG, inferior temporal gyrus (ITG), and the parahippocampal gyrus [23]. Wang et al. suggested that such PVA-related spontaneous activity may be related to visual memory consolidation processes [23]. Furthermore, PVA, as well as the occipito-parietal and occipito-temporal visual association areas, has been found to be activated by visual imagery tasks [24]. The occipital visual areas and frontal/parietal sensorimotor areas have been found to be concurrently activated during the visuo-spatial working memory task [25]. In this study, decreased ReHo values in the occipital lobe, temporal lobe and PoCG in T2DM patients may thus reflect decrement in either of these processes, possibly secondary to reduced function in the primary visual cortex. Notably, in the present study, both M+ and M− patients performed more poorly than nondiabetic controls on the TMT-A, TMT-B, TMT, and CFT-delayed recall tests, indicating that cognitive dysfunction in these patients are more likely to involve the domains of psychomotor speed, attention, executive function, and visual spatial skills respectively. These results are consistent with those from previous studies [10,14,26]. In addition, the ReHo values of the lingual gyrus/calcarine cortex were revealed to be associated with TMT-B, TMT and CFT-delayed scores in the diabetes group. Therefore, these results suggest that decreased ReHo values in the occipital lobe may imply the existence of cognitive impairment to some extent in T2DM patients. Our inference can be supported by previous rs-fMRI studies which showed that T2DM patients had significantly decreased ALFF and ReHo values [10,14] as well as decreased functional connectivity to the PCC [26] in

613

the occipital lobe compared to controls, which suggested the existence of impaired visual processing and visual spatial skills in T2DM patients. Besides the changes in the occipital lobe, decreased ReHo values were further observed in the temporal lobe (bilateral STG, MTG, and FG) in both M+ and M− patients compared with nondiabetic controls. In addition, ReHo values in the MTG were negatively correlated with TMT-B and TMT in T2DM patients. Since MTG is associated with verbal and visual semantic memory and working memory [27], our results may therefore reflect cognitive impairment in these aspects. Our inference can be supported by a study by Xia et al., who reported that T2DM patients had significantly decreased ALFF values in bilateral MTG, which is negatively associated with TMT-B performance [10]. In our study, ReHo values in STG were decreased in the T2DM group. Since STG is responsible for processing sounds, the decreased ReHo values in STG may suggest that T2DM patients potentially have hearing impairment, which is supported by previous studies, showing that T2DM patients were associated with an increased risk of age-related hearing impairment [28]. Significantly decreased ReHo values in PoCG in both M+ and M− T2DM patients were also found in this study. As the primary somatosensory cortex, PoCG is the main sensory receptive region for the sense of touch, proprioception, pain and temperature, which receives the bulk of the thalamocortical projections from the sensory input fields. Accordingly, decreased ReHo values in PoCG may partly contribute to a common diabetic limb pain and temperature sensation decline [29]. Decreased ReHo values can be observed in some brain areas as discussed above, increased ReHo values were also noted in some brain regions of both M+ and M− T2DM patients including bilateral PCu, SFG, MFG, and insula relative to controls, which may be interpreted as a compensatory mechanism for reduced cortical activity [30]. Compared with nondiabetic controls, both M+ and M− patients performed more poorly in TMT tests, which is known to be sensitive to frontal lobe damage [20]. Presumably, the ReHO values in the frontal lobe might be decreased in T2DM patients. However, in the present study, increased ReHo values in the frontal lobe were observed and no correlation between TMT scores and ReHo values in this area was detected in T2DM patients, which may be interpreted as a compensatory mechanism for the visual impairment in order to strengthen the control of external interference. The details of the compensation and association between TMT scores and ReHo values in the frontal lobe need to be further explored. Compared with the M− group, M+ patients showed decreased ReHo values in the left cuneus and SOG, indicating that M+ patients may have more decreased brain activity related to severely impaired function of visual processing and visual memory. There were no other ReHo differences in the brain between M+ and M− patients. Several epidemiological studies have shown that chronic hyperglycaemia and microangiopathy especially retinopathy contribute to cognitive decrement in both type 2 and type 1 diabetes [2,3]. Thereby, cognitive impairment would be more marked in patients who developed clinically measurable microangiopathy than in patients without microangiopathy. A previous study by van Duinkerken et al [11]. reported that MA− T1DM patients displayed increased connectivity in sensorimotor and secondary visual networks, whereas MA+ patients exhibited decreased connectivity in networks involved in working memory, attention, auditory and language processes, and motor and visual processing, which is partly consistent with our result. However, the difference in cognitive impairment between the two groups is less significant in our study than in van Duinkerken’s work. The possible reasons may lie in the differences in the selection criteria for the enrolled patients as well as the experiment design and rs-fMRI methods. The studies regarding the association between retinopathy and cognitive decline has

614

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615

shown that subjects with proliferative or laser-treated retinopathy tend to show greater cognitive decline [2,3,11]. Thus, in this respect we will focus on this more extreme group in the future study. Meanwhile, how and to what extent microangiopathy affects cognitive function in T2DM patients needs to be investigated intensively with other approaches such as independent component analysis (ICA) and functional connectivity (FC). The preliminary study has some limitations. First, this study had a relatively small sample size and was not a longitudinal research. Further study with more participants and following up on these patients are of great importance to evaluate whether the aberrant ReHo values could be markers for tracking the very early changes of brain function associated with T2DM. Second, besides diabetes retinopathy, other microvascular complications such as nephropathy and neuropathy may contribute to the results. Given the study design, we were not able to discard their relative contribution, which should be established in future studies. Third, in addition to microvascular disorders, other complications of T2DM such as hypertension and dyslipidemia may also contribute to the brain imaging abnormalities observed in this study. Therefore, additional studies should be performed to elucidate the exact relationship between the decreased ReHo values and cognition dysfunction in T2DM patients. Fourth, decreased stimulation of the primary visual cortex may result in impaired function of visual processing and visual memory. Although we have excluded the patients with severe visual or hearing loss that would affect the neuropsychological assessment and the ReHo analysis, we did not measure the patients’ visual acuity accurately and therefore could not adjust for it in the ReHo analysis and neuropsychological tests in the present cohort. Methods for measuring ReHo and neuropsychological changes in T2DM patients independent of visual acuity need to be considered in future studies. Since a combination of MRI methods such as the diffusion tensor imaging (DTI), cerebral blood flow (CBF), and cerebrovascular reactivity (CVR) techniques has been proved to improve the diagnosis accuracy, such combination method should be used in future studies to further assist our understanding of the pathophysiological mechanism underlying T2DM-induced cognitive impairment.

5. Conclusion The present study showed aberrant ReHo values in both M+ and M− T2DM patients, which were correlated with neuropsychological impairments in selected brain regions. Compared with the M− group, M+ patients showed decreased ReHo values in left cuneus and SOG, probably indicating more severely impaired function of visual processing and visual memory. The ReHo approach may be potentially valuable to investigate the pathophysiology of cognitive impairments in T2DM patients and detect the altered spontaneous brain activity prior to structural changes in these patients.

Conflict of interest The authors declare that there is no potential conflict of interests regarding the publication of this paper.

Acknowledgements This research was supported by the National Natural Science Foundation of China (81041050), National Key Clinical Specialties Construction Program of China, and health bureau of chongqing (2012-1-013 and 2012-1-017).

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejrad.2015.12. 024.

References [1] P.J. Spauwen, S. Köhler, F.R. Verhey, C.D. Stehouwer, M.P. van Boxtel, Effects of type 2 diabetes on 12-year cognitive change: results from the Maastricht Aging Study, Diabetes Care 36 (2013) 1554–1561. [2] A.M. Jacobson, C.M. Ryan, P.A. Cleary, B.H. Waberski, K. Weinger, G. Musen, et al., Diabetes Control and Complications Trial/EDIC Research Group. Biomedical risk factors for decreased cognitive functioning in type 1 diabetes: an 18 year follow-up of the Diabetes Control and Complications Trial (DCCT) cohort, Diabetologia 54 (2011) 245–255. [3] J. Ding, M.W. Strachan, R.M. Reynolds, B.M. Frier, I.J. Deary, F.G. Fowkes, et al., Edinburgh Type 2 Diabetes Study Investigators. Diabetic retinopathy and cognitive decline in older people with type 2 diabetes: the Edinburgh Type 2 Diabetes Study, Diabetes 59 (2010) 2883–2889. [4] N. Patton, T. Aslam, T. Macgillivray, A. Pattie, I.J. Deary, B. Dhillon, Retinal vascular image analysis as a potential screening tool for cerebrovascular disease: a rationale based on homology between cerebral and retinal microvasculatures, J. Anat. 206 (2005) 319–348. [5] L. Ni, R. Qi, L.J. Zhang, J. Zhong, G. Zheng, X. Wu, et al., Brain regional homogeneity changes following transjugular intrahepatic portosystemic shunt in cirrhotic patients support cerebral adaptability theory—a resting-state functional MRI study, Eur. J. Radiol. 83 (2014) 578–583. [6] T. Wu, X. Long, Y. Zang, L. Wang, M. Hallett, K. Li, et al., Regional homogeneity changes in patients with Parkinson’s disease, Hum. Brain Mapp. 30 (2009) 1502–1510. [7] Q.Z. Wu, D.M. Li, W.H. Kuang, T.J. Zhang, S. Lui, X.Q. Huang, Abnormal regional spontaneous neural activity in treatment-refractory depression revealed by resting-state fMRI, Hum. Brain Mapp. 32 (2011) 1290–1299. [8] Z. Zhang, Y. Liu, T. Jiang, B. Zhou, N. An, H. Dai, Altered spontaneous activity in Alzheimer’s disease and mild cognitive impairment revealed by regional homogeneity, Neuroimage 59 (2012) 1429–1440. [9] G. Musen, A.M. Jacobson, N.R. Bolo, D.C. Simonson, M.E. Shenton, R.L. McCartney, et al., Resting-state brain functional connectivity is altered in type 2 diabetes, Diabetes 61 (2012) 2375–2379. [10] W. Xia, S. Wang, Z. Sun, F. Bai, Y. Zhou, P. Wang, et al., Altered baseline brain activity in type 2 diabetes: a resting-state fMRI study, Psychoneuroendocrinology 38 (2013) 2493–2501. [11] E. van Duinkerken, M.M. Schoonheim, E.J. Sanz-Arigita, R.G. IJzerman, A.C. Moll, F.J. Snoek, 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 61 (2012) 1814–1821. [12] Y. Zang, T. Jiang, Y. Lu, Y. He, L. Tian, Regional homogeneity approach to fMRI data analysis, Neuroimage 22 (2004) 394–400. [13] W. Fang, F. Lv, T. Luo, O. Cheng, W. Liao, K. Sheng, et al., Abnormal regional homogeneity in patients with essential tremor revealed by resting-state functional MRI, PLoS One 8 (2013) e69199. [14] Y. Cui, Y. Jiao, Y.C. Chen, K. Wang, B. Gao, S. Wen, et al., Altered spontaneous brain activity in type 2 diabetes: a resting-state functional MRI study, Diabetes 63 (2014) 749–760. [15] American Diabetes Association, Diagnosis and classification of diabetes mellitus, Diabetes Care 36 (Suppl. 1) (2013) S67–S74. [16] S.J. Aldington, E.M. Kohner, S. Meuer, R. Klein, A.K. Sjølie, Methodology for retinal photography and assessment of diabetic retinopathy: the EURODIAB IDDM complications study, Diabetologia 38 (1995) 437–444. [17] A.M. Berhane, E.J. Weil, W.C. Knowler, R.G. Nelson, R.L. Hanson, Albuminuria and estimated glomerular filtration rate as predictors of diabetic end-stage renal disease and death, Clin. J. Am. Soc. Nephrol. 6 (2011) 2444–2451. [18] American Diabetes AssociationA.J. Bouhon, A.I. Vinik, J.C. Arezzo, V. Bril, E.L. Feldman, R. Freeman, et al., Diabetic neuropathies: a statement by the American Diabetes Association, Diabetes Care 28 (2005) 956–962. [19] E.L. Helkala, M. Kivipelto, M. Hallikainen, K. Alhainen, H. Heinonen, J. Tuomilehto, et al., Usefullness of repeated presentation of Mini-Mental State Examination as a diagnostic procedure- a population-based study, Acta Neurol. Scand. 106 (2002) 341–346. [20] A. Ble, S. Volpato, G. Zuliani, J.M. Guralnik, S. Bandinelli, F. Lauretani, et al., Executive function correlates with walking speed in older persons: the InCHIANTI study, J. Am. Geriatr. Soc. 53 (2005) 410–415. [21] M. Kendall, J.D. Gibbons, Rank Correlation Methods, Oxford University Press, Oxford, 1990. [22] M.E. Raichle, A.M. MacLeod, A.Z. Snyder, W.J. Powers, D.A. Gusnard, G.L. Shulman, A default mode of brain function, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 676–682. [23] K. Wang, T. Jiang, C. Yu, L. Tian, J. Li, Y. Liu, et al., Spontaneous activity associated with primary visual cortex: a resting-state FMRI study, Cereb. Cortex 18 (2008) 697–704.

J. Peng et al. / European Journal of Radiology 85 (2016) 607–615 [24] S. Lambert, E. Sampaio, Y. Mauss, C. Scheiber, Blindness and brain plasticity: contribution of mental imagery? An fMRI study, Brain Res. Cogn. Brain Res. 20 (2004) 1–11. [25] H. Kwon, A.L. Reiss, V. Menon, Neural basis of protracted developmental changes in visuo-spatial working memory, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 13336–13341. [26] Y.C. Chen, Y. Jiao, Y. Cui, S.A. Shang, J. Ding, Y. Feng, et al., Aberrant brain functional connectivity related to insulin resistance in type 2 diabetes: a resting-state fMRI study, Diabetes Care 37 (2014) 1689–1696. [27] H.T. van Schie, A.A. Wijers, R.B. Mars, J.S. Benjamins, L.A. Stowe, Processing of visual semantic information to concrete words: temporal dynamics and neural mechanisms indicated by event-related brain potentials, Cogn. Neuropsychol. 22 (2005) 364–386.

615

[28] K.E. Bainbridge, H.J. Hoffman, C.C. Cowie, Risk factors for hearing impairment among U.S. adults with diabetes: National Health and Nutrition Examination Survey 1999–2004, Diabetes Care 34 (2011) 1540–1545. [29] F. Cauda, K. Sacco, S. Duca, D. Cocito, F. D’Agata, G.C. Geminiani, et al., Altered resting state in diabetic neuropathic pain, PLoS One 4 (2009) e4542. [30] S.D. Roosendaal, M.M. Schoonheim, H.E. Hulst, E.J. Sanz-Arigita, S.M. Smith, J.J. Geurts, et al., Resting state networks change in clinically isolated syndrome, Brain 133 (2010) 1612–1621.