White matter microstructural variability mediates the relation between obesity and cognition in healthy adults

NeuroImage 172 (2018) 239–249

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/neuroimage

White matter microstructural variability mediates the relation between obesity and cognition in healthy adults Rui Zhang (张睿)a, b, Frauke Beyer a, b, Leonie Lampe a, c, Tobias Luck c, d, Steffi G. Riedel-Heller c, d, Markus Loeffler c, e, Matthias L. Schroeter a, f, Michael Stumvoll b, g, Arno Villringer a, b, c, f, A. Veronica Witte a, b, * a

Max Planck Institute for Human Cognitive and Brain Sciences, 04103 Leipzig, Germany Collaborative Research Centre 1052 ’Obesity Mechanisms', Subproject A1, Faculty of Medicine, Leipzig University, 04103 Leipzig, Germany c Leipzig Research Center for Civilization Diseases (LIFE), Leipzig University, 04103 Leipzig, Germany d Institute of Social Medicine, Occupational Health and Public Health (ISAP), Leipzig University, 04103 Leipzig, Germany e Institute for Medical Informatics, Statistics, and Epidemiology (IMSE), Leipzig University, 04107 Leipzig, Germany f Clinic of Cognitive Neurology, Leipzig University Hospital, 04103 Leipzig, Germany g Medical Department, Division of Endocrinology and Nephrology, Leipzig University Hospital, 04103 Leipzig, Germany b

A R T I C L E I N F O

A B S T R A C T

Keywords: DTI Fractional anisotropy Executive functions Processing speed

Obesity has been linked with structural and functional brain changes. However, the impact of obesity on brain and cognition in aging remains debatable, especially for white matter. We therefore aimed to determine the effects of obesity on white matter microstructure and potential implications for cognition in a well-characterized large cohort of healthy adults. In total, 1255 participants (50% females, 19–80 years, BMI 16.8–50.2 kg/m2) with diffusion-weighted magnetic resonance imaging at 3T were analysed. Tract-based spatial statistics (TBSS) probed whether body mass index (BMI) and waist-to-hip ratio (WHR) were related to fractional anisotropy (FA). We conducted partial correlations and mediation analyses to explore whether obesity or regional FA were related to cognitive performance. Analyses were adjusted for demographic, genetic, and obesity-associated confounders. Results showed that higher BMI and higher WHR were associated with lower FA in multiple white matter tracts (p < 0.05, FWE-corrected). Mediation analyses provided evidence for indirect negative effects of higher BMI and higher WHR on executive functions and processing speed through lower FA in fiber tracts connecting (pre)frontal, visual, and associative areas (indirect paths, jßj  0.01; 99% jCIj > 0). This large cross-sectional study showed that obesity is correlated with lower FA in multiple white matter tracts in otherwise healthy adults, independent of confounders. Moreover, although effect sizes were small, mediation results indicated that visceral obesity was linked to poorer executive functions and lower processing speed through lower FA in callosal and associative fiber tracts. Longitudinal studies are needed to support this hypothesis.

Introduction Numerous epidemiological studies showed that obesity, most often measured using body mass index (BMI), is associated with accelerated cognitive decline and an increased risk for dementia later in life (Fergenbaum et al., 2009; Fitzpatrick et al., 2009; Gustafson et al., 2009; Whitmer et al., 2008), for reviews see (Albanese et al., 2017; Emmerzaal et al., 2015; Prickett et al., 2015). However, this detrimental relationship has been questioned: For example, Qizilbash et al. (2015) analysed retrospectively online health-care data of over 2 million people in the UK and reported that obese people had a lower dementia risk. Others

observed that obesity in midlife but not in old age exerts detrimental effects on cognitive decline (Gustafson, 2015). These phenomena might be explained by reverse causation due to weight loss at the preclinical stage of dementia, a hypothesis supported recently by data including over 1.3 million participants from 39 prospective cohorts (Kivim€aki et al., 2017). Clarifying the debate if obesity exerts negative effects on brain health even in older age is crucial from a public health perspective. Because in 2016, about 40% of adults worldwide were overweight, and the prevalence rate of both obesity and Alzheimer's disease is rapidly increasing (Alzheimer’s Association, 2016; WHO, 2016). Cognitive decline has been linked to neuroimaging markers of grey

* Corresponding author. Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences, Stephanstrasse 1a, 04103 Leipzig, Germany. E-mail address: [email protected] (A.V. Witte). https://doi.org/10.1016/j.neuroimage.2018.01.028 Received 14 September 2017; Received in revised form 5 December 2017; Accepted 12 January 2018 1053-8119/© 2018 Published by Elsevier Inc.

R. Zhang et al.

NeuroImage 172 (2018) 239–249

One third of the DWI scans were rated as non-usable due to chemical shift artifacts identified by visual inspection on raw diffusion-weighted and diffusion tensor V1 images (for an example see supplement Fig. S1). Such an artifact is caused by insufficient fat suppression during image acquisition, and incomplete fat suppression could still occur because it highly depends on the suppression method and scanner-specific settings (Tax et al., 2016). We excluded those scans because this artifact would bias tensor estimation locally, and correction during preprocessing is difficult (Tax et al., 2016). Of 1713 usable DWI scans, we excluded participants due to brain pathologies (n ¼ 169), cancer treatment (n ¼ 59), and intake of medication that affects the central nervous system (n ¼ 156) (Fig. 1). We further excluded participants with missing values (n ¼ 60) in all cognitive tests or BMI. To exclude severe cognitive impairments, we excluded participants with an average global cognitive sum-score (see below) below 3 standard deviations of the subgroup mean (n ¼ 6 < 60y, and n ¼ 8 60–80y) (Masouleh et al., 2016), leaving 1255 participants for main analysis.

matter atrophy (Karas et al., 2004) and regional white matter damage (Bartzokis et al., 2004; Brickman et al., 2006; Lampe et al., 2017; Madden et al., 2012; Vernooij et al., 2009). Obesity could cause white matter damage due to multiple physiological factors, including high blood pressure, low-grade inflammation, and insulin resistance (Pugazhenthi et al., 2016; Shefer et al., 2013; Willette et al., 2015). Thus, obesity-associated regional white matter damage might be one underlying mechanism of cognitive differences in the general population, either independent of age or as a trigger or booster of age-associated changes. This hypothesis, however, has not yet been systematically explored. Regarding obesity and global measures of white matter, available neuroimaging studies yielded inconclusive results (Bobb et al., 2014; Driscoll et al., 2012; Karlsson et al., 2013; Raji et al., 2010; Walther et al., 2010). Using studies implementing sensitive diffusion tensor imaging (DTI) to assess fractional anisotropy (FA), a measure of directional diffusion of water molecules in the white matter (Alexander et al., 2007), correlated obesity in healthy adults to lower FA with relative consistency (Bettcher et al., 2013; Karlsson et al., 2013; Kullmann et al., 2016; Mueller et al., 2011; Ryan and Walther, 2014; Stanek et al., 2011; Verstynen et al., 2012). These obesity-associated differences were found in diverse, not necessarily overlapping, brain regions and were partially restricted to either men or women. For the first time, a recent study reported a positive relationship between obesity and FA (Birdsill et al., 2017). Thus, it is still unclear if obesity is detrimental for white matter microstructure, and if this effect is age-, sex- or region-specific. In addition, if obesity-associated white matter microstructural changes would relate to cognitive changes remains unclear. Previous conflicting findings might be explained by differences among studies in methodology but also when considering specific aspects of obesity and demographic confounders (Kullmann et al., 2015). For example, several studies reported detrimental effects for visceral obesity (often measured by waist-to-hip ratio, WHR), but overall obesity as measured by BMI showed less or no effect on white matter microstructure (Gustafson et al., 2009; Karlsson et al., 2013; Whitmer et al., 2008). In addition, previous studies often neglected the possibility of residual confounding, e.g., through risk factors related to obesity (hypertension, Rahmouni, 2014; or diabetes, Van Gaal et al., 2006), white matter lesions (Kim et al., 2017), and/or cognitive decline (e.g., education, Garibotto et al., 2008) and common polymorphisms in the Apolipoprotein E (APOE) ε4-gene (Ryan et al., 2011). This highlights the importance of a well-characterized sample, tight control of confounders, and sensitive cognitive testing. We therefore aimed to determine if obesity is correlated with regional white matter microstructure in a large cohort of otherwise healthy men and women across the adult lifespan by evaluating linear and nonlinear effects of both BMI and WHR, applying strict exclusion criteria, and controlling possible confounders. In addition, we tested if associations between obesity and regional white matter microstructure would translate into inter-individual differences in main domains of cognitive performance using structural equation path analyses (Hayes and Preacher, 2014), a method frequently used to explore mediating pathways between age, metabolic factors, brain structure, and cognition (Kerti et al., 2013; Mander et al., 2013; Ryan and Walther, 2014). We hypothesized that obesity was associated with lower FA in both men and women and with cognitive functions through lower FA in areas that are supportive of executive control, memory, and processing speed in older individuals.

Image acquisition and processing Diffusion-weighted images were acquired on a 3-T MRI scanner (Verio, Siemens, Germany) with a double spin-echo encoding sequence which lasted 16 min 8 s (repetition time [TR] ¼ 13800 ms; echo time [TE] ¼ 100 ms; field of view [FOV] ¼ 220  220  123 mm3; matrix ¼ 128  128; voxel size ¼ 1.7  1.7  1.7 mm3; maximum bvalue ¼ 1000 s/mm2 in 60 directions, and 7 volumes with b-value ¼ 0 s/ mm2). Participants were scanned from 8am to 3pm on weekdays. Following the ENGIMA (Enhancing Neuroimaging Genetics Through Meta Analysis, www.enigma.ini.usc.edu)-DTI protocol, we used the Diffusion Toolbox, which is part of FMRIB Software Library (FSL v5.0.9) (Smith et al., 2004), to process diffusion images. Firstly, the raw data for each participant were corrected for head motion and distortions by registering them to the no diffusion-weighted (b ¼ 0) image with an affine transformation. Subsequently, non-brain structures were removed by the BET tool (Smith, 2002) with b ¼ 0 image as reference. FA images were then created by fitting a tensor model to the raw diffusion data with rotated bvec files. After that, tract-based spatial statistics (TBSS) were used to carry out voxelwise statistical analysis (Smith et al., 2006). Briefly, all participants' FA images were first registered to a FMRIB58_FA standard-space image with a non-linear registration tool FNIRT (Andersson et al., 2007a, 2007b). This step created a standard-space version of each individual's FA image and merged all into a single 4D image. From the 4D image, a mean FA image was generated and thinned to create a mean FA skeleton. The skeleton was then used for projecting each subject's aligned FA data and creating the 4D skeletonized FA image for voxelwise cross-subject statistics. The voxelwise analyses were carried out for all voxels with FA  0.3 to exclude extreme cross-subject variability within small individual white matter tracts as well as incorrect alignments in the registration. Cognitive evaluation We assessed cognitive performance using the neuropsychological test battery of the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) (Morris et al., 1989). These tests comprised the Trail-Making Test (TMT) parts A and B, Verbal Fluency Test ‘Animals’ and ‘S’ words, and a 10-word list for the verbal episodic memory test. For the latter, “learning” was defined as the sum of correctly named words out of three consecutive learning trials. “Recall” was defined as the number of correctly recalled words from the list after a delay, in which participants performed a figure copy task (~5 min), and “recognition” was defined as the sum of correctly recognized words from a list of 20 mixed words presented after the recall trial. Three composite scores for the cognitive domains of executive functions, memory performance, and processing speed were defined according to previous studies (Masouleh et al., 2016; van de Rest et al., 2008; Witte et al., 2014) as: ‘executive functions’ ¼ [z

Materials and methods Participants The study participants were from the population-based “Leipzig Research Center for Civilization Diseases” (LIFE-Adult) cohort in Leipzig, Germany (Loeffler et al., 2015). A total of 2637 participants underwent magnetic resonance imaging (MRI) of the head, among which 2523 participants completed a diffusion-weighted imaging (DWI) protocol. 240

R. Zhang et al.

NeuroImage 172 (2018) 239–249

Fig. 1. Flowchart of the study. From 1713 participants with usable diffusion-weighted scans, n ¼ 1255 entered whole-brain voxelwise analyses of obesity, namely body mass index (BMI, n ¼ 1255) and waist-to-hip ratio (WHR, n ¼ 1251), on white matter microstructure. Subsequently, obesity and regional white matter microstructural variability were related to cognitive performance using correlations and mediation analyses. Abbreviations: DWI, diffusionweighted imaging; SD, standard deviation; TBSS, tract-based spatial statistics; FA, fractional anisotropy; APOE, Apolipoprotein E.

241

R. Zhang et al.

NeuroImage 172 (2018) 239–249

phonemic fluency þ z semantic fluency - z TMT(part B - part A)/part A]/3; ‘memory performance’ ¼ (z learning þ z recall þ z recognition)/3; ‘processing speed’ ¼ -z TMT part A. See supplementary Fig. S2 for distribution of the composite scores and Table S1 for overview of the raw cognitive subtest-scores. In addition, a global cognitive sum-score was defined as the average age-adjusted performance in executive functions, memory performance and processing speed in a young group (<60 y) and in an older group (60–80 y) (Masouleh et al., 2016).

analyses controlled for age and sex. Therefore, we a-priori selected the following major white matter tracts based on previous results linking white matter microstructure with cognitive performance (Borghesani et al., 2013; Gold et al., 2012; Madden et al., 2009a, 2009b; Ryan and Walther, 2014; Salami et al., 2012): uncinate fasciculus, corpus callosum, anterior thalamic radiation, superior and inferior longitudinal fasciculus, and cingulum. These regions were masked on the white matter skeleton with the JHU ICBM-DTI-81 atlas, and individual mean FA was extracted in regional clusters that were significantly associated with BMI according to TBSS analyses (FWE-corrected p < 0.05) (see Fig. 2). Subsequently, we analysed associations of mean FA in these BMI/WHR-associated clusters with cognitive performance in the mediation analyses. The indirect effect was considered significant if zero was not included in the 99% bias-corrected confidence intervals (CI) based on 10,000 bootstrap samples. We controlled for education, APOE-ε4 status, and comorbidities, namely hypertension and diabetes, as confounders in a subsequent model.

Anthropometric assessments and confounder definition BMI (kg/m2) was calculated as weight in kilograms divided by height squared in meters, both measured on a standard scale to the nearest 0.01 kg or 0.1 cm. WHR was defined as waist circumference divided by hip circumference measured by an ergonomic circumference measuring tape (SECA 201) to the nearest 0.1 cm. Education was evaluated based on the International Standard Classification of Education (Unesco, 1997), and years of education were counted as years one had been at school according to the respective graduation level (self-reported). Hypertension was defined as a systolic blood pressure higher than 140 mmHg or a diastolic blood pressure higher than 90 mmHg (Mancia et al., 2013), by indirect assumption when they took antihypertensive medication, or by clinical diagnosis. Diabetes mellitus was defined by self-reported diagnosis and use of medication. APOE genotype was used dichotomously: APOE-ε4 carrier versus non-carrier. Individual white matter hyperintensity volume (WMHV) was calculated with an automatic segmentation tool (LesionTOADS) (Lampe et al., 2017; Shiee et al., 2010) and adjusted for global white matter volume.

Missing values Because of missing WHR measures in 4 subjects, 1251 participants were included in the voxelwise cross-subject whole-brain analysis for the association between WHR and FA. Missing values of WMHV were replaced by the median WMHV value of the respective Fazekas’ scale category (Fazekas et al., 1987), or, if the Fazekas rating was not available, by the median of the respective age group (<60 y or 60–80y). Each cognitive composite score was calculated based on the average of the remaining available subtests when there were missing values in any of the cognitive subtests (Masouleh et al., 2016). Participants missing one of the composite scores were excluded in further analyses with pairwise strategy. A log transformation was performed if the absolute value of skewness exceeded one. All analyses were conducted with psych and Hmisc package in R (Harrell et al., 2016; R Core Team, 2016; Revelle, 2017) except for the mediation analysis which was done by PROCESS SPSS macro (Hayes and Preacher, 2014) in SPSS 24 (PASW, SPSS, IBM).

Statistical analyses Obesity and white matter To test the association between obesity and white matter microstructure, we performed voxelwise cross-subject whole-brain analyses of BMI and WHR, respectively, on FA skeletonized maps, adjusting for age, sex, and WMHV. Nonparametric inference testing was conducted using the “randomise” tool in FSL with 10,000 permutations (Winkler et al., 2014). In addition, we tested for quadratic effects of obesity measures and for interactions between obesity measures and age or sex. In adjusted models we controlled for potential confounding effects of obesity-related comorbidities (hypertension and diabetes), and APOE-ε4 status. Further sensitivity analyses were done in males and females separately. Multiple comparisons were corrected by threshold-free cluster enhancement (Smith and Nichols, 2009) and family-wise error (FWE) at a significance level of p < 0.05. Effect sizes of these analyses were compared using a one-sample t-test of the voxelwise differences in t-values.

Results We included 1255 non-demented participants in the main analysis; see Table 1 for an overview of anthropometric characteristics. Obesity and white matter microstructure TBSS analysis detected a significant association between higher BMI and lower FA in multiple, mainly bilateral regions of the cerebral white matter tracts (Fig. 3, p < 0.05, FWE-corrected, controlled for age, sex, and WMHV). Results remained stable when correcting additionally for confounding effects of diabetes, hypertension, and APOE-ε4 carrier status (Fig. S3). The affected white matter tracts included the entire corpus callosum, cerebral peduncle, internal capsule, corona radiata, superior and inferior longitudinal fasciculus, anterior and posterior thalamic radiation, external capsule, superior cerebellar peduncle, and uncinate fasciculus. Considering WHR as a predictor, we detected an overall very slightly larger effect (voxel-wise comparison, Cohen's d WHR>BMI ¼ 0.08,

Cognitive performance Relations between BMI or WHR and cognitive composite scores were examined with partial correlations. Significance level was set at an adjusted p-value ¼ 0.017 for multiple comparisons < 0.05 (Bonferronicorrected p < 0.05, 3 domains tested, 2-sided). To examine the potential role of regional BMI- and WHR-associated white matter variability on the relation between BMI or WHR and cognitive performance (Ryan and Walther, 2014; Scholz et al., 2009), we performed simple mediation

Fig. 2. Regions of interest on the white matter skeleton: corpus callosum genu (CC-G), body (CC-B) and splenium (CC-S), inferior and superior longitudinal fasciculus (ILF and SLF), cingulum (CG), uncinate fasciculus (UNC) and anterior corona radiata (ACR).

242

R. Zhang et al.

NeuroImage 172 (2018) 239–249

Cognitive performance analyses

Table 1 Characteristics of included study participants (n ¼ 1255). Characteristic

Descriptives

Sex (female), n (%) Age, mean (SD), range, y BMI, mean (SD), range, kg/m2 WHRa, mean (SD), range Years of Educationb, mean (SD), range, y APOE-ε4 carrierc, n (%) Hypertension, n (%) Diabetesd, n(%) Executive functione, mean (SD), range Processing speed (log10)f, mean (SD), range Memory performance (log10)g, mean (SD), range

636 (50.7) 55.43 (16.05), 19–80 25.82 (3.69), 16.81–50.15 0.91 (0.09), 0.64–1.17 10.7 (2.1), 0–13 276 (23.5) 701 (55.9) 97 (8.3) 0.01 (0.72), 3.27–3.03 0.47 (0.13), 0–0.88 0.51 (0.13), 0–0.83

We found a significant negative correlation between WHR and memory performance (r ¼ 0.09, p ¼ 0.005, controlled for age and sex), however the effect diminished after further adjustments, namely education, diabetes, hypertension, and APOE-ε4 (Table 2). There were no significant correlations between BMI and cognition after adjusting for confounders (all p > 0.03, Table 2). Mediation analyses While the direct link between obesity and cognitive performance appeared to be limited, simple mediation analyses indicated specific causal pathways between obesity, regional white matter microstructure, and cognitive function. That is, higher BMI and higher WHR contributed to poorer performance in executive functions and processing speed through lower FA in clusters within the superior and inferior lateral fasciculus, left anterior thalamic radiation, and in the body of the corpus callosum (99% jCIj > 0, Figs. 6 and 7, Tables 3 and 4). Associations remained stable for WHR-related effects after adjusting for all confounders in the corpus callosum, left arcuate thalamic radiation, and left superior longitudinal fasciculus. We did not observe robust associations for further mediating pathways.

Abbreviations: SD, standard deviation; BMI, body mass index; WHR, waist-to-hip ratio; APOE, Apolipoprotein E. a n ¼ 1251. b n ¼ 1188. c n ¼ 1175. d n ¼ 1165. e n ¼ 1193. f n ¼ 1188. g n ¼ 1206.

p < 104) which was a significant association between higher WHR and lower FA on multiple white matter tracts almost throughout the whole white matter skeleton in the frontal, temporal, and parietal lobes (Fig. 4, p < 0.05, FWE-corrected, controlled for age, sex, and WMHV). Results remained stable when correcting additionally for confounding effects of diabetes, hypertension, and APOE-ε4 carrier status (Fig. S4). There were neither quadratic effects nor main interaction effects of BMI or WHR by age or sex in this large dataset after appropriate correction (FWE-corrected, p > 0.05). However, sensitivity analyses for sex groups separately indicated that the association was slightly stronger in female (BMI: voxel-wise comparison, Cohen's d female>male ¼ 0.05, p < 104, Fig. 5; WHR: voxel-wise comparison, Cohen's d fema4 le>male ¼ 0.34, p < 10 ). We did not observe significant positive associations between BMI or WHR and FA. As head motion can affect data quality and was found to be correlated with BMI (Beyer et al., 2017; Hodgson et al., 2017), we additionally controlled for head motion, measured by mean framewise displacement (FD) derived by rigid-body co-registration during preprocessing (Power et al., 2012). This did not attenuate the results (Fig. S5–6). In addition, excluding participants with mean FD > 1 mm gave similar results (data not shown).

Discussion In this large cross-sectional study of 1255 older individuals, we found independent associations between obesity and lower FA in multiple white matter tracts comprising the corpus callosum, superior and inferior longitudinal fasciculus, corona radiata, thalamic radiation, uncinate fasciculus, cingulum, and brainstem. While the direct link between obesity and cognition in this cohort was limited, mediation analyses supported the existence of detrimental pathways linking obesity to worse executive functions and lower processing speed through lower FA in callosal and associative fiber tracts. Obesity and white matter Our finding of an inverse association between obesity and white matter FA extends previous small studies (all n < 60) in young and middle-aged participants that reported negative correlations in few white matter areas (Karlsson et al., 2013; Mueller et al., 2011; Shott et al., 2015; Verstynen et al., 2013; Xu et al., 2013; Yau et al., 2014), for a review see (Kullmann et al., 2015). Two recent studies conducted in medium-sized

Fig. 3. Association of higher body mass index (BMI) and lower white matter fractional anisotropy (FA) adjusted for age, sex, and volume of white matter hyperintensities (p < 0.05, family-wise-error-corrected, n ¼ 1225). This figure is layered on a T1-weighted image (coordinates according to MNI152 template). Color bar represents t-value. Abbreviations: L, left; R, right. 243

R. Zhang et al.

NeuroImage 172 (2018) 239–249

Fig. 4. Association of higher waist-to-hip ratio (WHR) and lower white matter fractional anisotropy (FA) adjusted for age, sex, and volume of white matter hyperintensities (p < 0.05, family-wise-error-corrected, n ¼ 1251). This figure is layered on a T1-weighted image (coordinates according to MNI152 template). Color bar represents t-value. Abbreviations: L, left; R, right.

Fig. 5. Inverse correlation between body mass index (BMI) and fractional anisotropy (FA) in the white matter between sex groups. First row is the correlation in men and the second row is the correlation in women, showing lower FA associated with higher BMI adjusted for age and volume of white matter hyperintensities (p < 0.05, family-wise-error-corrected). This figure is layered on a T1-weighted image (coordinates according to MNI152 template). Color bar represents t-value. Abbreviations: L, left; R, right.

Table 2 Correlations between BMI or WHR and cognitive performance. Executive function

Memory performance

Processing speed

n

r (95%CI)

p

n

r (95%CI)

p

n

r (95%CI)

p

Model 1a BMI WHR

1193 1189

0.06 (0.12, 0.007) 0.06 (0.12, 0.004)

0.03 0.03

1206 1202

0.04 (0.09, 0.02) ¡0.09 (¡0.14, ¡0.03)

0.19 0.005*

1188 1184

0.02 (0.04, 0.08)

0.51 >0.9

Model 2b BMI WHR

1089 1085

0.03 (0.09, 0.03)

>0.9 0.26

1048 1044

0.006 (0.07, 0.05) 0.07 (0.13, 0.008)

0.85 0.026

1086 1082

0.009 (0.05, 0.07) 0.02 (0.07, 0.04)

0.76 0.6

Abbreviations: BMI, body mass index; WHR, waist-to-hip ratio. Adjusted p ¼ 0.017. a Controlled age and sex as confounders. b Model 1 added education, APOE-ε4 status and comorbidities (diabetes, hypertension) as confounders.

samples (n ¼ 103, lifespan sample, Stanek et al., 2011; n ¼ 238, middle-aged hypertensive patients, Papageorgiou et al., 2017) extended results that obese and/or overweight participants had partially lower FA compared to their lean counterparts in the corpus callosum, thalamic radiation, uncinate fasciculus, and inferior and superior longitudinal fasciculus. This is in line with the current findings, yet without showing such a widespread pattern of regions associated with continuous

increases in BMI and WHR. The affected regions observed in our sample were also partially reported when examining obesity-associated white matter microstructure specifically in older groups with sample sizes between 15 and 138 subjects (Bettcher et al., 2013; Bolzenius et al., 2015; Marks et al., 2010; Ryan and Walther, 2014), yet again showing a smaller regional correlations between BMI and FA. Besides the smaller sample sizes and related lower statistical power, this could also be due to 244

R. Zhang et al.

NeuroImage 172 (2018) 239–249

In addition, previous studies suggested sex-specific effects of obesity on FA (Mueller et al., 2011). This could be due to known anatomical differences or hormonal effects (Liu et al., 2010; Menzler et al., 2011; Peper et al., 2011). Exploratory comparisons of whole-brain results from analyses separately in males and females in our study gave hints for a somewhat stronger association in female compared to male, specifically for WHR. One explanation for these observations might be a stronger collinearity of WHR with age in males explaining twice as much of the variance compared to females (age ~ WHR R2 males ¼ 0.41, R2 females ¼ 0.19): This might have led to a certain overfitting in the male group. In addition, an interaction of obesity measures on FA by sex did not show significant effects in the pooled analysis. Thus, our results do not provide strong evidence for the existence of sex differences in the effects of obesity on FA. Future studies incorporating direct measurements of body composition could help to further understand differences between BMI, WHR, and the role of visceral fat on FA, as well as potential differences in men and women. Contrary to our hypotheses, we could not observe nonlinear trends for obesity effects, i.e., a quadratic term of BMI or WHR would independently explain more variance in the data than BMI or WHR as a linear term. In addition, there was no significant age-by-BMI or age-by-WHR interaction when additionally controlling for linear effects of age and BMI/WHR. This implies that effects of obesity across the lifespan are rather linear. However, the recruitment of participants for the LIFE-study was intended to focus on participants over 60 years of age (please see Loeffler et al., 2015, for details) and our mean age was 55.43 years  16.05 SD, which might have led to a certain underrepresentation of young adults in our cohort. In addition, due to existing collinearities between higher age and higher BMI/WHR, we cannot completely exclude nonlinear effects of obesity over the lifespan.

Fig. 6. Clusters from body mass index (BMI)-associated white matter tracts extracted within the left inferior longitudinal fasciculus (ILF) and within the right superior longitudinal fasciculus (SLF) (threshold-free cluster enhancement and family-wise error corrected p < 0.05).

methodological limitations, for example, atlas-based region-of-interest approach to delineate white matter differences in these studies, compared to the more sensitive whole-brain voxel-wise TBSS analysis in our study. Our results further indicate a slightly stronger effect of WHR compared to BMI. As WHR is an indicator of higher abdominal fat and better reflects visceral fat than BMI (Shuster et al., 2012), this could imply that abdominal fat deposition, in particular visceral fat, is more detrimental to white matter FA variability than overall fat or body weight. When measuring body composition using e.g., bioelectrical impedance analysis, women have significantly higher fat percentages than men, even at the same BMI (Gallagher et al., 1996; Kyle et al., 2003).

Obesity and cognitive performance In our cohort, correlations between obesity and cognition appeared to be weak or non-significant when adjusted for possible confounding effects of age, sex and education. This might be insightful given that effect sizes for unique effects of obesity on cognitive decline were estimated to be rather small (Livingston et al., 2017). In addition, our cohort comprised of cognitively healthy adults, combining relatively strict exclusion criteria with a wide age range that could already account for a Fig. 7. Clusters from waist-to-hip-ratio (WHR)-associated white matter tracts extracted within the corpus callosum body (CC-B), right inferior longitudinal fasciculus (ILF), left anterior corona radiata (ACR) and superior longitudinal fasciculus (SLF) (threshold-free cluster enhancement and family-wise error corrected p < 0.05).

245

R. Zhang et al.

NeuroImage 172 (2018) 239–249

Table 3 Mediation effect of regional FA on the association between BMI and cognitive function. SLF R executive functions

SLF R processing speed

ILF L processing speed

β

p or 99.5% CI

β

p or 99.5% CI

β

p or 99.5% CI

0.06 0.08 0.13 ¡0.01 0.06

0.028 0.008 <0.001 [¡0.025, ¡0.0005] 0.06

0.02 0.08 0.07 ¡0.005 0.02

0.5 0.007 0.008 [¡0.025, ¡0.0001] 0.38

0.02 0.08 0.07 ¡0.005 0.02

0.5 0.007 0.006 [¡0.017, ¡0.0001] 0.38

a

Model 1 Total effect c (BMI on cognition) a (BMI on FA) b (FA on cognition) Mediation effect a*b (BMI on cognition via FA) Direct effect c’ (BMI on cognition)

Abbreviations: BMI, body mass index; FA, fractional anisotropy, SLF, superior longitudinal fasciculus, ILF, inferior longitudinal fasciculus, R, right, L, left. a Controlled age and sex as confounders.

Table 4 Mediation effect of regional FA on the association between WHR and cognitive function.

Model 1a Total effect c (WHR on cognition) a (WHR on FA) b (FA on cognition) Mediation effect a*b (WHR on cognition via FA) Direct effect c’ (WHR on cognition) Model 2b Total effect c (WHR on cognition) a (WHR on FA) b (FA on cognition) Mediation effect a*b (WHR on cognition via FA) Direct effect c’ (WHR on cognition)

CC body executive functions

SLF R executive functions

SLF R processing speed

SLF L executive functions

SLF L processing speed

β

p or 99.5% CI

β

p or 99% CI

β

p or 99% CI

β

p or 99% CI

β

p or 99% CI

0.09

0.035

0.09

0.035

0.005

0.8

0.09

0.035

0.005

0.89

0.13 0.08 ¡0.01

0.11 0.13 ¡0.02

0.004 <0.001 [¡0.037, ¡0.0023] 0.08

0.12 0.08 ¡0.01

0.08

0.002 0.001 [¡0.032, ¡0.002] 0.069

0.13 0.09 ¡0.01

0.01

0.003 0.002 [¡0.025, ¡0.001] 0.8

0.12 0.10 ¡0.01

0.08

<0.001 0.015 [¡0.026, ¡0.0008] 0.061

0.006

0.002 <0.001 [¡0.027, ¡0.002] 0.86

0.05

0.26

0.02

0.6

0.05

0.26

0.02

0.59

0.11 0.08 ¡0.009

0.008 0.014 [¡0.025, ¡0.0002] 0.36

0.11 0.09 ¡0.01

0.012 <0.0001 [¡0.026, ¡0.0005] 0.8

0.11 0.09 ¡0.01

0.009 0.002 [¡0.031, ¡0.0003] 0.38

0.11 0.09 ¡0.01

0.008 <0.001 [¡0.027, ¡0.0009] 0.80

0.04

0.08

n.s.

0.01

0.04

0.009

ACR L executive functions

ACR L processing speed

ILF R executive functions

ILF R processing speed

β

p or 99.5% CI

β

p or 99% CI

β

p or 99% CI

β

p or 99% CI

0.09 0.12 0.09 ¡0.01

0.035 <0.001 0.015 [¡0.027, ¡0.0001] 0.062

0.001 0.12 0.09 ¡0.01

0.9 <0.001 0.003 [¡0.027, ¡0.002] 0.8

0.09 0.10 0.10 ¡0.01

0.035 0.014 0.001 [¡0.029, ¡0.0003] 0.06

0.005 0.11 0.08 ¡0.01

0.9 0.008 0.002 [¡0.024, ¡0.0008] 0.9

0.26 0.014 0.006 [¡0.026, ¡0.0002] 0.36

0.02 0.09 0.11 ¡0.01

a

Model 1 Total effect c (WHR on cognition) a (WHR on FA) b (FA on cognition) Mediation effect a*b (WHR on cognition via FA) Direct effect c’ (WHR on cognition)

0.08

0.01

0.08

0.004

b

Model 2 Total effect c (WHR on cognition) a (WHR on FA) b (FA on cognition) Mediation effect a*b (WHR on cognition via FA) Direct effect c’ (WHR on cognition)

0.05 0.09 0.10 ¡0.009 0.04

0.6 0.014 <0.001 [¡0.027, ¡0.0006] 0.8

0.01

n.s.

n.s.

Abbreviations: WHR, waist-hip ratio; FA, fractional anisotropy; CC, corpus callosum; SLF, superior longitudinal fasciculus; ACR, anterior thalamic radiation; ILF, inferior longitudinal fasciculus; R, right; L, left. a Controlled age and sex as confounders. b Model 1 added education, APOE-ε4 status and comorbidities (diabetes, hypertension) as confounders.

callosum, the anterior thalamic radiation, and the superior and inferior longitudinal fasciculus (Boorman et al., 2009; Borghesani et al., 2013; Jacobs et al., 2013; Kennedy and Raz, 2009; Kochunov et al., 2009; Salthouse, 2011; partly even independent of age; Brickman et al., 2006; Jokinen et al., 2007). Associations in the corpus callosum, left anterior thalamic radiation, and left superior longitudinal fasciculus with the visceral obesity marker WHR remained stable even after controlling for education, diabetes, and hypertension, as well as genetic factors underscoring the robustness of these effects. We confirmed and further extended previous results of 95 older females, showing mediating effects

large amount of variability in cognitive performance. While our cross-sectional dataset is unable to prove causality, mediation analyses provided initial evidence based on a large amount of observations for the existence of a detrimental pathway between obesity, white matter microstructure, and cognitive performance. Accordingly, mediation analyses indicated that obesity contributed to poorer executive functions and lower processing speed through lower FA in obesityassociated regions of the white matter skeleton, independent of age and sex. Those were located within major tracts which have previously been linked to higher-order cognitive abilities, including the corpus

246

R. Zhang et al.

NeuroImage 172 (2018) 239–249

longitudinal studies on obesity, white matter changes, and cognition are warranted to confirm our hypotheses and to identify factors that help to prevent or reverse the development of obesity and obesity-associated brain damage.

of obesity on cognition through white matter FA (Ryan and Walther, 2014). In the latter study, diffusion measures in mostly temporal and parietal white matter also predicted memory performance (Ryan and Walther, 2014), a result that was not confirmed by our data. However note that the fornix region as a major tract of the limbic system implicated in memory processes did not show associations with BMI or WHR at the chosen significance level and was therefore not considered in the cognitive mediation analyses. Regarding the underlying mechanisms, FA has been discussed as a measure for white matter microstructure due to its high sensitivity to neuropathology and microstructural architecture (Alexander et al., 2007). Thus, lower FA in white matter tracts might indicate dysfunctional properties of connecting axonal fibers, thereby provoking disadvantages in cognitive processing (Johansen-Berg and Behrens, 2006). Obesity-associated changes in white matter FA might be due to low-grade inflammation, a condition frequently related to obesity. For example, in healthy populations, increased inflammatory markers such as interleukin-6 and C-reactive protein have been associated with reduced white matter FA (Gianaros et al., 2013; Shott et al., 2015). Consistently, diffusion measures in overweight and obese individuals were predicted by plasma fibrinogen levels, a marker of inflammation (Cazettes et al., 2011). These cytokines are produced by adipose tissue (in particular by visceral fat; Fontana et al., 2007; Johnson et al., 2012) and can enter the central nervous system as a result of a disrupted blood–brain barrier, possibly due to maintaining an energy dense Western diet when being obese (Hargrave et al., 2016; Johnson et al., 2012; Pugazhenthi et al., 2016). Future longitudinal studies investigating the role of obesity-associated inflammation in provoking white matter damage and related cognitive decline are needed to further strengthen these hypotheses.

Funding This work was supported by grants of the European Union, the European Regional Development Fund, the Free State of Saxony within the framework of the excellence initiative, the LIFE–Leipzig Research Center for Civilization Diseases, University of Leipzig [project numbers: 713241202, 14505/2470, 14575/2470], and the German Research Foundation [CRC 1052 “Obesity mechanisms”, Project A1, A. Villringer/M. Stumvoll and WI 3342/3-1, A.V. Witte]. Financial disclosures The authors declare no conflict of interest. Acknowledgments The authors would like to thank all participants and the staff at the LIFE study center. The authors are also grateful to Alfred Anwander, J€ oran Lepsien, Toralf Mildner, Harald M€ oller, Karsten Müller, Andre Pampel, and Robert Trampel for methodological input and helpful discussions. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi. org/10.1016/j.neuroimage.2018.01.028.

Limitations

References

Our study has some limitations. First, due to the cross-sessional design, no causal relationships can be established based on our results. However, path analyses indicated mediating links between obesity, white matter FA, and cognitive performance. Regarding the memory performance, the CERAD-recognition memory subtest showed some ceiling effects in our cohort, which might have affected statistical power. Limitations shared by all DTI and voxel-based studies include potential registration errors and spatial smoothing. However, we increased the mean FA skeleton threshold to 0.3 to exclude tracts that were not well aligned across subjects (Smith et al., 2006). Another limitation of our study is that we did not evaluate axial and radial diffusivity (AD and RD), further DTI measures which might give additional insights into white matter microstructure (Alexander et al., 2007). Strengths of the study include the large sample size of community-dwelling adults with extensive data on demographic, cognitive assessments, and potential confounding factors. Although evaluation of education was somewhat skewed to higher education, our cohort showed the full range from 0 to 13 years of education. Note that participants were randomly invited from the Leipzig city registry, thus the study sample should well represent the general population of Eastern Germany. In addition, participants underwent high resolution DWI at 3T with 60 directions which guaranteed robust estimates of tensor-derived properties (Jones et al., 2013).

Albanese, E., Launer, L.J., Egger, M., Prince, M.J., Giannakopoulos, P., Wolters, F.J., Egan, K., 2017. Body mass index in midlife and dementia: systematic review and meta-regression analysis of 589,649 men and women followed in longitudinal studies. Alzheimers. Dement 8, 165–178. https://doi.org/10.1016/j.dadm.2017.05. 007. Alexander, A.L., Lee, J.E., Lazar, M., Field, A.S., 2007. Diffusion tensor imaging of the brain. Neurotherapeutics, Advances in Neuroimaging/Neuroethics 4, 316–329. https://doi.org/10.1016/j.nurt.2007.05.011. Alzheimer’s Association, 2016. 2016 Alzheimer's disease facts and figures. Alzheimers. Dement 12, 459–509. https://doi.org/10.1016/j.jalz.2016.03.001. Andersson, J.L.R., Jenkinson, M., Smith, S., 2007a. Non-linear Registration, aka Spatial Normalisation. FMRIB technical report TR07JA2. Andersson, J.L.R., Jenkinson, M., Smith, S., 2007b. Non-linear Optimisation. FMRIB technical report TR07JA1. Bartzokis, G., Sultzer, D., Lu, P.H., Nuechterlein, K.H., Mintz, J., Cummings, J.L., 2004. Heterogeneous age-related breakdown of white matter structural integrity: implications for cortical “disconnection” in aging and Alzheimer's disease. Neurobiol. Aging 25, 843–851. https://doi.org/10.1016/j.neurobiolaging.2003.09.005. Bettcher, B.M., Walsh, C.M., Watson, C., Miller, J.W., Green, R., Patel, N., Miller, B.L., Neuhaus, J., Yaffe, K., Kramer, J.H., 2013. Body mass and white matter integrity: the influence of vascular and inflammatory markers. PLoS One 8, e77741. https://doi. org/10.1371/journal.pone.0077741. Beyer, F., Kharabian Masouleh, S., Huntenburg, J.M., Lampe, L., Luck, T., RiedelHeller, S.G., Loeffler, M., Schroeter, M.L., Stumvoll, M., Villringer, A., Witte, A.V., 2017. Higher body mass index is associated with reduced posterior default mode connectivity in older adults. Hum. Brain Mapp. 38, 3502–3515. https://doi.org/10. 1002/hbm.23605. Birdsill, A.C., Oleson, S., Kaur, S., Pasha, E., Ireton, A., Tanaka, H., Haley, A., 2017. Abdominal obesity and white matter microstructure in midlife. Hum. Brain Mapp. 38, 3337–3344. https://doi.org/10.1002/hbm.23576. Bobb, J.F., Schwartz, B.S., Davatzikos, C., Caffo, B., 2014. Cross-sectional and longitudinal association of body mass index and brain volume. Hum. Brain Mapp. 35, 75–88. https://doi.org/10.1002/hbm.22159. Bolzenius, J.D., Laidlaw, D.H., Cabeen, R.P., Conturo, T.E., McMichael, A.R., Lane, E.M., Heaps, J.M., Salminen, L.E., Baker, L.M., Scott, S.E., Cooley, S.A., Gunstad, J., Paul, R.H., 2015. Brain structure and cognitive correlates of body mass index in healthy older adults. Behav. Brain Res. 278, 342–347. https://doi.org/10.1016/j.bbr. 2014.10. Boorman, E.D., Behrens, T.E.J., Woolrich, M.W., Rushworth, M.F.S., 2009. How Green is the grass on the other Side? Frontopolar cortex and the evidence in favor of

Conclusions We found that obesity in adults is associated with lower FA in multiple white matter tracts throughout the brain. Moreover, although effect sizes were small, mediation results indicated that visceral obesity was linked to poorer executive functions and lower processing speed through lower FA in callosal and associative fiber tracts. Thus, our findings might help to assess the complex interplay between obesity and subtle cognitive impairments, and point to the need for curative options. Upcoming

247

R. Zhang et al.

NeuroImage 172 (2018) 239–249 Jones, D.K., Kn€ osche, T.R., Turner, R., 2013. White matter integrity, fiber count, and other fallacies: the do's and don'ts of diffusion MRI. Neuroimage 73, 239–254. https://doi.org/10.1016/j.neuroimage.2012.06.081. Karas, G.B., Scheltens, P., Rombouts, S.A.R.B., Visser, P.J., van Schijndel, R.A., Fox, N.C., Barkhof, F., 2004. Global and local gray matter loss in mild cognitive impairment and Alzheimer's disease. Neuroimage 23, 708–716. https://doi.org/10.1016/j. neuroimage.2004.07.006. Karlsson, H.K., Tuulari, J.J., Hirvonen, J., Lepom€aki, V., Parkkola, R., Hiltunen, J., Hannukainen, J.C., Soinio, M., Pham, T., Salminen, P., Nuutila, P., Nummenmaa, L., 2013. Obesity is associated with white matter atrophy: a combined diffusion tensor imaging and voxel-based morphometric study. Obesity 21, 2530–2537. https://doi. org/10.1002/oby.20386. Kennedy, K.M., Raz, N., 2009. Aging white matter and cognition: differential effects of regional variations in diffusion properties on memory, executive functions, and speed. Neuropsychologia 47, 916–927. https://doi.org/10.1016/j.neuropsychologia. 2009.01.001. Kerti, L., Witte, A.V., Winkler, A., Grittner, U., Rujescu, D., Fl€ oel, A., 2013. Higher glucose levels associated with lower memory and reduced hippocampal microstructure. Neurology 81, 1746–1752. https://doi.org/10.1212/01.wnl.0000435561.00234.ee. Kim, K.W., Seo, H., Kwak, M.-S., Kim, D., 2017. Visceral obesity is associated with white matter hyperintensity and lacunar infarct. Int. J. Obes. 41, 683–688. https://doi.org/ 10.1038/ijo.2017.13. Kivim€aki, M., Luukkonen, R., Batty, G.D., Ferrie, J.E., Pentti, J., Nyberg, S.T., Shipley, M.J., Alfredsson, L., Fransson, E.I., Goldberg, M., Knutsson, A., Koskenvuo, M., Kuosma, E., Nordin, M., Suominen, S.B., Theorell, T., Vuoksimaa, E., Westerholm, P., Westerlund, H., Zins, M., Kivipelto, M., Vahtera, J., Kaprio, J., SinghManoux, A., Jokela, M., 2017. Body mass index and risk of dementia: analysis of individual-level data from 1.3 million individuals. Alzheimers. Dement. https://doi. org/10.1016/j.jalz.2017.09.016. Kochunov, P., Robin, D.A., Royall, D.R., Coyle, T., Lancaster, J., Kochunov, V., Schlosser, A.E., Fox, P.T., 2009. Can structural MRI indices of cerebral integrity track cognitive trends in executive control function during normal maturation and adulthood? Hum. Brain Mapp. 30, 2581–2594. https://doi.org/10.1002/hbm.20689. Kullmann, S., Callaghan, M.F., Heni, M., Weiskopf, N., Scheffler, K., H€aring, H.-U., Fritsche, A., Veit, R., Preissl, H., 2016. Specific white matter tissue microstructure changes associated with obesity. Neuroimage 125, 36–44. https://doi.org/10.1016/j. neuroimage.2015.10.006. Kullmann, S., Schweizer, F., Veit, R., Fritsche, A., Preissl, H., 2015. Compromised white matter integrity in obesity. Obes. Rev. 16, 273–281. https://doi.org/10.1111/obr. 12248. Kyle, U.G., Schutz, Y., Dupertuis, Y.M., Pichard, C., 2003. Body composition interpretation: contributions of the fat-free mass index and the body fat mass index. Nutrition 19, 597–604. https://doi.org/10.1016/S0899-9007(03)00061-3. Lampe, L., Kharabian-Masouleh, S., Kynast, J., Arelin, K., Steele, C.J., L€ offler, M., Witte, A.V., Schroeter, M.L., Villringer, A., Bazin, P.-L., 2017. Lesion location matters: the relationships between white matter hyperintensities on cognition in the healthy elderly. J. Cereb. Blood Flow Metab 1–8. https://doi.org/10.1177/ 0271678X17740501. Liu, F., Vidarsson, L., Winter, J.D., Tran, H., Kassner, A., 2010. Sex differences in the human corpus callosum microstructure: a combined T2 myelin-water and diffusion tensor magnetic resonance imaging study. Brain Res. 1343, 37–45. https://doi.org/ 10.1016/j.brainres.2010.04.064. Livingston, G., Sommerlad, A., Orgeta, V., Costafreda, S.G., Huntley, J., Ames, D., Ballard, C., Banerjee, S., Burns, A., Cohen-Mansfield, J., Cooper, C., Fox, N., Gitlin, L.N., Howard, R., Kales, H.C., Larson, E.B., Ritchie, K., Rockwood, K., Sampson, E.L., Samus, Q., Schneider, L.S., Selbæk, G., Teri, L., Mukadam, N., 2017. Dementia prevention, intervention, and care. Lancet 390, 2673–2734. https://doi. org/10.1016/S0140-6736(17). Loeffler, M., Engel, C., Ahnert, P., Alfermann, D., Arelin, K., Baber, R., Beutner, F., Binder, H., Br€ahler, E., Burkhardt, R., Ceglarek, U., Enzenbach, C., Fuchs, M., Glaesmer, H., Girlich, F., Hagendorff, A., H€antzsch, M., Hegerl, U., Henger, S., Hensch, T., Hinz, A., Holzendorf, V., Husser, D., Kersting, A., Kiel, A., Kirsten, T., Kratzsch, J., Krohn, K., Luck, T., Melzer, S., Netto, J., Nüchter, M., Raschpichler, M., Rauscher, F.G., Riedel-Heller, S.G., Sander, C., Scholz, M., Sch€ onknecht, P., Schroeter, M.L., Simon, J.-C., Speer, R., St€aker, J., Stein, R., St€ obel-Richter, Y., Stumvoll, M., Tarnok, A., Teren, A., Teupser, D., Then, F.S., T€ onjes, A., Treudler, R., Villringer, A., Weissgerber, A., Wiedemann, P., Zachariae, S., Wirkner, K., Thiery, J., 2015. The LIFE-Adult-Study: objectives and design of a population-based cohort study with 10,000 deeply phenotyped adults in Germany. BMC Publ. Health 15, 691. https://doi.org/10.1186/s12889-015-1983-z. Madden, D.J., Bennett, I.J., Burzynska, A., Potter, G.G., Chen, N.-K., Song, A.W., 2012. Diffusion tensor imaging of cerebral white matter integrity in cognitive aging. Biochimica et Biophysica Acta (BBA) - molecular Basis of Disease. Imaging Brain Aging and Neurodegenerative Disease 1822, 386–400. https://doi.org/10.1016/j. bbadis.2011.08.003. Madden, D.J., Bennett, I.J., Song, A.W., 2009a. Cerebral white matter integrity and cognitive aging: contributions from diffusion tensor imaging. Neuropsychol. Rev. 19, 415–435. https://doi.org/10.1007/s11065-009-9113-2. Madden, D.J., Spaniol, J., Costello, M.C., Bucur, B., White, L.E., Cabeza, R., Davis, S.W., Dennis, N.A., Provenzale, J.M., Huettel, S.A., 2009b. Cerebral white matter integrity mediates adult age differences in cognitive performance. J. Cogn. Neurosci 21, 289–302. https://doi.org/10.1162/jocn.2009.21047. Mancia, G., Fagard, R., Narkiewicz, K., Red on, J., Zanchetti, A., B€ ohm, M., Christiaens, T., Cifkova, R., De Backer, G., Dominiczak, A., Galderisi, M., Grobbee, D.E., Jaarsma, T., Kirchhof, P., Kjeldsen, S.E., Laurent, S., Manolis, A.J., Nilsson, P.M., Ruilope, L.M.,

alternative courses of action. Neuron 62, 733–743. https://doi.org/10.1016/j. neuron.2009.05.014. Borghesani, P.R., Madhyastha, T.M., Aylward, E.H., Reiter, M.A., Swarny, B.R., Schaie, K.W., Willis, S.L., 2013. The association between higher order abilities, processing speed, and age are variably mediated by white matter integrity during typical aging. Neuropsychologia 51, 1435–1444. https://doi.org/10.1016/j. neuropsychologia.2013.03.005. Brickman, A.M., Zimmerman, M.E., Paul, R.H., Grieve, S.M., Tate, D.F., Cohen, R.A., Williams, L.M., Clark, C.R., Gordon, E., 2006. Regional white matter and neuropsychological functioning across the adult lifespan. Biol. Psychiatry 60, 444–453. https://doi.org/10.1016/j.biopsych.2006.01.011. Cazettes, F., Cohen, J.I., Yau, P.L., Talbot, H., Convit, A., 2011. Obesity-mediated inflammation may damage the brain circuit that regulates food intake. Brain Res. 1373, 101–109. https://doi.org/10.1016/j.brainres.2010.12.008. Driscoll, I., Beydoun, M.A., An, Y., Davatzikos, C., Ferrucci, L., Zonderman, A.B., Resnick, S.M., 2012. Midlife obesity and trajectories of brain volume changes in older adults. Hum. Brain Mapp. 33, 2204–2210. https://doi.org/10.1002/hbm.21353. Emmerzaal, T.L., Kiliaan, A.J., Gustafson, D.R., 2015. 2003-2013: a decade of body mass index, Alzheimer's disease, and dementia. J. Alzheimers. Dis 43, 739–755. https:// doi.org/10.3233/JAD-141086. Fazekas, F., Chawluk, J., Alavi, A., Hurtig, H., Zimmerman, R., 1987. MR signal abnormalities at 1.5 T in Alzheimer's dementia and normal aging. Am. J. Roentgenol. 149, 351–356. https://doi.org/10.2214/ajr.149.2.351. Fergenbaum, J.H., Bruce, S., Lou, W., Hanley, A.J.G., Greenwood, C., Young, T.K., 2009. Obesity and lowered cognitive performance in a Canadian First Nations population. Obesity 17, 1957–1963. https://doi.org/10.1038/oby.2009.161. Fitzpatrick, A.L., Kuller, L.H., Lopez, O.L., Diehr, P., O'Meara, E.S., Longstreth Jr., W.T., Luchsinger, J.A., 2009. Midlife and late-life obesity and the risk of dementia: cardiovascular health study. Arch. Neurol. 66, 336–342. https://doi.org/10.1001/ archneurol.2008.582. Fontana, L., Eagon, J.C., Trujillo, M.E., Scherer, P.E., Klein, S., 2007. Visceral fat adipokine secretion is associated with systemic inflammation in obese humans. Diabetes 56, 1010–1013. https://doi.org/10.2337/db06-1656. Gallagher, D., Visser, M., Sepúlveda, D., Pierson, R.N., Harris, T., Heymsfield, S.B., 1996. How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? Am. J. Epidemiol. 143, 228–239. Garibotto, V., Borroni, B., Kalbe, E., Herholz, K., Salmon, E., Holtoff, V., Sorbi, S., Cappa, S.F., Padovani, A., Fazio, F., Perani, D., 2008. Education and occupation as proxies for reserve in aMCI converters and AD: FDG-PET evidence. Neurology 71, 1342–1349. https://doi.org/10.1212/01.wnl.0000327670.62378.c0. Gianaros, P.J., Marsland, A.L., Sheu, L.K., Erickson, K.I., Verstynen, T.D., 2013. Inflammatory pathways link socioeconomic inequalities to white matter architecture. Cerebr. Cortex 23, 2058–2071. https://doi.org/10.1093/cercor/bhs191. Gold, B.T., Johnson, N.F., Powell, D.K., Smith, C.D., 2012. White matter integrity and vulnerability to Alzheimer's disease: preliminary findings and future directions. Biochim. Biophys. Acta 1822, 416–422. https://doi.org/10.1016/j.bbadis.2011.07. 009. Gustafson, D., 2015. BMI and dementia: feast or famine for the brain? The Lancet Diabetes & Endocrinology 3, 397–398. https://doi.org/10.1016/s2213-8587(15) 00085-6. Gustafson, D.R., B€ ackman, K., Waern, M., Ostling, S., Guo, X., Zandi, P., Mielke, M.M., Bengtsson, C., Skoog, I., 2009. Adiposity indicators and dementia over 32 years in Sweden. Neurology 73, 1559–1566. https://doi.org/10.1212/WNL. 0b013e3181c0d4b6. Hargrave, S.L., Jones, S., Davidson, T.L., 2016. The outward spiral: a vicious cycle model of obesity and cognitive dysfunction. Curr. Opin. Behav. Sci. 9, 40–46. https://doi. org/10.1016/j.cobeha.2015.12.001. Harrell Jr., F.E., 2016. Package “Hmisc” [WWW Document]. URL. https://cran.r-project. org/web/packages/Hmisc/Hmisc.pdf. Hayes, A.F., Preacher, K.J., 2014. Statistical mediation analysis with a multicategorical independent variable. Br. J. Math. Stat. Psychol. 67, 451–470. https://doi.org/10. 1111/bmsp.12028. Hodgson, K., Carless, M.A., Kulkarni, H., Curran, J.E., Sprooten, E., Knowles, E.E., Mathias, S., G€ oring, H.H.H., Yao, N., Olvera, R.L., Fox, P.T., Almasy, L., Duggirala, R., Blangero, J., Glahn, D.C., 2017. Epigenetic age acceleration assessed with human white-matter images. J. Neurosci. 0177–0117 https://doi.org/10.1523/JNEUROSCI. 0177-17.2017. Jacobs, H.I.L., Leritz, E.C., Williams, V.J., Van Boxtel, M.P.J., Elst, W., van der, Jolles, J., Verhey, F.R.J., McGlinchey, R.E., Milberg, W.P., Salat, D.H., 2013. Association between white matter microstructure, executive functions, and processing speed in older adults: the impact of vascular health. Hum. Brain Mapp. 34, 77–95. https://doi. org/10.1002/hbm.21412. Johansen-Berg, H., Behrens, T.E.J., 2006. Just pretty pictures? What diffusion tractography can add in clinical neuroscience. Curr. Opin. Neurol. 19, 379–385. https://doi.org/10.1097/01.wco.0000236618.82086.01. Johnson, A.R., Justin Milner, J., Makowski, L., 2012. The inflammation highway: metabolism accelerates inflammatory traffic in obesity. Immunol. Rev. 249, 218–238. https://doi.org/10.1111/j.1600-065X.2012.01151.x. Jokinen, H., Ryberg, C., Kalska, H., Ylikoski, R., Rostrup, E., Stegmann, M.B., Waldemar, G., Madureira, S., Ferro, J.M., Straaten, E.C.W., van, Scheltens, P., Barkhof, F., Fazekas, F., Schmidt, R., Carlucci, G., Pantoni, L., Inzitari, D., Erkinjuntti, T., 2007. Corpus callosum atrophy is associated with mental slowing and executive deficits in subjects with age-related white matter hyperintensities: the LADIS Study. J. Neurol. Neurosurg. Psychiatry 78, 491–496. https://doi.org/10. 1136/jnnp.2006.096792.

248

R. Zhang et al.

NeuroImage 172 (2018) 239–249 Shiee, N., Bazin, P.-L., Ozturk, A., Reich, D.S., Calabresi, P.A., Pham, D.L., 2010. A topology-preserving approach to the segmentation of brain images with multiple sclerosis lesions. Neuroimage 49, 1524–1535. https://doi.org/10.1016/j. neuroimage.2009.09.005. Shott, M.E., Cornier, M.-A., Mittal, V.A., Pryor, T.L., Orr, J.M., Brown, M.S., Frank, G.K.W., 2015. Orbitofrontal cortex volume and brain reward response in obesity. Int. J. Obes. 39, 214–221. https://doi.org/10.1038/ijo.2014.121. Shuster, A., Patlas, M., Pinthus, J.H., Mourtzakis, M., 2012. The clinical importance of visceral adiposity: a critical review of methods for visceral adipose tissue analysis. Br. J. Radiol. 85, 1–10. https://doi.org/10.1259/bjr/38447238. Smith, S.M., 2002. Fast robust automated brain extraction. Hum. Brain Mapp. 17, 143–155. https://doi.org/10.1002/hbm.10062. Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.E., Watkins, K.E., Ciccarelli, O., Cader, M.Z., Matthews, P.M., Behrens, T.E.J., 2006. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31, 1487–1505. https://doi.org/10.1016/j.neuroimage.2006.02.024. Smith, S.M., Jenkinson, M., Woolrich, M.W., Beckmann, C.F., Behrens, T.E.J., JohansenBerg, H., Bannister, P.R., De Luca, M., Drobnjak, I., Flitney, D.E., Niazy, R.K., Saunders, J., Vickers, J., Zhang, Y., De Stefano, N., Brady, J.M., Matthews, P.M., 2004. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23 (Suppl. 1), S208–S219. https://doi.org/10.1016/j. neuroimage.2004.07.051. Smith, S.M., Nichols, T.E., 2009. Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence and localisation in cluster inference. Neuroimage 44, 83–98. https://doi.org/10.1016/j.neuroimage.2008.03.061. Stanek, K.M., Grieve, S.M., Brickman, A.M., Korgaonkar, M.S., Paul, R.H., Cohen, R.A., Gunstad, J.J., 2011. Obesity is associated with reduced white matter integrity in otherwise healthy adults. Obesity 19, 500–504. https://doi.org/10.1038/oby.2010. 312. Tax, C.M.W., Vos, S.B., Leemans, A., 2016. Checking and correcting DTI data. In: Diffusion Tensor Imaging. Springer, New York, NY, pp. 127–150. https://doi.org/10. 1007/978-1-4939-3118-7_7. Unesco, E., 1997. International Standard Classification of Education: ISCED 1997. UNESCO. van de Rest, O., Geleijnse, J.M., Kok, F.J., van Staveren, W.A., Dullemeijer, C., Olderikkert, M.G.M., Beekman, A.T.F., de Groot, C.P.G.M., 2008. Effect of fish oil on cognitive performance in older subjects: a randomized, controlled trial. Neurology 71, 430–438. https://doi.org/10.1212/01.wnl.0000324268.45138.86. Van Gaal, L.F., Mertens, I.L., De Block, C.E., 2006. Mechanisms linking obesity with cardiovascular disease. Nature 444, 875–880. https://doi.org/10.1038/nature05487. Vernooij, M.W., Ikram, M.A., Vrooman, H.A., Wielopolski, P.A., Krestin, G.P., Hofman, A., Niessen, W.J., Van der Lugt, A., Breteler, M.M.B., 2009. White matter microstructural integrity and cognitive function in a general elderly population. Arch. Gen. Psychiatr. 66, 545. https://doi.org/10.1001/archgenpsychiatry.2009.5. Verstynen, T.D., Weinstein, A., Erickson, K.I., Sheu, L.K., Marsland, A.L., Gianaros, P.J., 2013. Competing physiological pathways link individual differences in weight and abdominal adiposity to white matter microstructure. Neuroimage 79, 129–137. https://doi.org/10.1016/j.neuroimage.2013.04.075. Verstynen, T.D., Weinstein, A.M., Schneider, W.W., Jakicic, J.M., Rofey, D.L., Erickson, K.I., 2012. Increased body mass index is associated with a global and distributed decrease in white matter microstructural integrity. Psychosom. Med. 74, 682–690. https://doi.org/10.1097/PSY.0b013e318261909c. Walther, K., Birdsill, A.C., Glisky, E.L., Ryan, L., 2010. Structural brain differences and cognitive functioning related to body mass index in older females. Hum. Brain Mapp. 31, 1052–1064. https://doi.org/10.1002/hbm.20916. Whitmer, R.A., Gustafson, D.R., Barrett-Connor, E., Haan, M.N., Gunderson, E.P., Yaffe, K., 2008. Central obesity and increased risk of dementia more than three decades later. Neurology 71, 1057–1064. https://doi.org/10.1212/01.wnl. 0000306313.89165.ef. WHO, 2016. WHO j Obesity and Overweight. Willette, A.A., Johnson, S.C., Birdsill, A.C., Sager, M.A., Christian, B., Baker, L.D., Craft, S., Oh, J., Statz, E., Hermann, B.P., Jonaitis, E.M., Koscik, R.L., La Rue, A., Asthana, S., Bendlin, B.B., 2015. Insulin resistance predicts brain amyloid deposition in late middle-aged adults. Alzheimers. Dement 11, 504–510 e1. https://doi.org/10. 1016/j.jalz.2014.03.011. Winkler, A.M., Ridgway, G.R., Webster, M.A., Smith, S.M., Nichols, T.E., 2014. Permutation inference for the general linear model. Neuroimage 92, 381–397. https://doi.org/10.1016/j.neuroimage.2014.01.060. Witte, A.V., Kerti, L., Hermannst€adter, H.M., Fiebach, J.B., Schreiber, S.J., Schuchardt, J.P., Hahn, A., Fl€ oel, A., 2014. Long-chain Omega-3 fatty acids improve brain function and structure in older adults. Cerebr. Cortex 24, 3059–3068. https:// doi.org/10.1093/cercor/bht163. Xu, J., Li, Y., Lin, H., Sinha, R., Potenza, M.N., 2013. Body mass index correlates negatively with white matter integrity in the fornix and corpus callosum: a diffusion tensor imaging study. Hum. Brain Mapp. 34, 1044–1052. https://doi.org/10.1002/ hbm.21491. Yau, P.L., Kang, E.H., Javier, D.C., Convit, A., 2014. Preliminary evidence of cognitive and brain abnormalities in uncomplicated adolescent obesity. Obesity 22, 1865–1871. https://doi.org/10.1002/oby.20801.

Schmieder, R.E., Sirnes, P.A., Sleight, P., Viigimaa, M., Waeber, B., Zannad, F., Task Force Members, 2013. 2013 ESH/ESC guidelines for the management of arterial hypertension: the task force for the management of arterial hypertension of the European society of hypertension (ESH) and of the European society of cardiology (ESC). J. Hypertens. 31, 1281–1357. https://doi.org/10.1097/01.hjh.0000431740. 32696.cc. Mander, B.A., Rao, V., Lu, B., Saletin, J.M., Lindquist, J.R., Ancoli-Israel, S., Jagust, W., Walker, M.P., 2013. Prefrontal atrophy, disrupted NREM slow waves and impaired hippocampal-dependent memory in aging. Nat. Neurosci. 16, 357–364. https://doi. org/10.1038/nn.3324. Marks, B.L., Katz, L.M., Styner, M., Smith, J.K., 2010. Aerobic fitness and obesity: relationship to cerebral white matter integrity in the brain of active and sedentary older adults. Br. J. Sports Med. 1208–1215. Masouleh, S.K., Arelin, K., Horstmann, A., Lampe, L., Kipping, J.A., Luck, T., RiedelHeller, S.G., Schroeter, M.L., Stumvoll, M., Villringer, A., Witte, A.V., 2016. Higher body mass index in older adults is associated with lower gray matter volume: implications for memory performance. Neurobiol. Aging 40, 1–10. https://doi.org/ 10.1016/j.neurobiolaging.2015.12.020. Menzler, K., Belke, M., Wehrmann, E., Krakow, K., Lengler, U., Jansen, A., Hamer, H.M., Oertel, W.H., Rosenow, F., Knake, S., 2011. Men and women are different: diffusion tensor imaging reveals sexual dimorphism in the microstructure of the thalamus, corpus callosum and cingulum. Neuroimage 54, 2557–2562. https://doi.org/10. 1016/j.neuroimage.2010.11.029. Morris, J.C., Heyman, A., Mohs, R.C., Hughes, J.P., van Belle, G., Fillenbaum, G., Mellits, E.D., Clark, C., 1989. The Consortium to establish a registry for Alzheimer's disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer's disease. Neurology 39, 1159–1165. Mueller, K., Anwander, A., M€ oller, H.E., Horstmann, A., Lepsien, J., Busse, F., Mohammadi, S., Schroeter, M.L., Stumvoll, M., Villringer, A., Pleger, B., 2011. Sexdependent influences of obesity on cerebral white matter investigated by diffusiontensor imaging. PLoS One 6, e18544. https://doi.org/10.1371/journal.pone. 0018544. Papageorgiou, I., Astrakas, L.G., Xydis, V., Alexiou, G.A., Bargiotas, P., Tzarouchi, L., Zikou, A.K., Kiortsis, D.N., Argyropoulou, M.I., 2017. Abnormalities of brain neural circuits related to obesity: a Diffusion Tensor Imaging study. Magn. Reson. Imaging 37, 116–121. https://doi.org/10.1016/j.mri.2016.11.018. Peper, J.S., van den Heuvel, M.P., Mandl, R.C.W., Pol, H.E.H., van Honk, J., 2011. Sex steroids and connectivity in the human brain: a review of neuroimaging studies. Psychoneuroendocrinology 36, 1101–1113. https://doi.org/10.1016/j.psyneuen. 2011.05.004. Power, J.D., Barnes, K.A., Snyder, A.Z., Schlaggar, B.L., Petersen, S.E., 2012. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage 59, 2142–2154. https://doi.org/10.1016/j.neuroimage.2011. 10.018. Prickett, C., Brennan, L., Stolwyk, R., 2015. Examining the relationship between obesity and cognitive function: a systematic literature review. Obes. Res. Clin. Pract. 9, 93–113. https://doi.org/10.1016/j.orcp.2014.05.001. Pugazhenthi, S., Qin, L., Reddy, P.H., 2016. Common neurodegenerative pathways in obesity, diabetes, and Alzheimer's disease. Biochim. Biophys. Acta (BBA) - Mol. Basis Dis. 1863, 1037–1045. https://doi.org/10.1016/j.bbadis.2016.04.017. Qizilbash, N., Gregson, J., Johnson, M.E., Pearce, N., Douglas, I., Wing, K., Evans, S.J.W., Pocock, S.J., 2015. BMI and risk of dementia in two million people over two decades: a retrospective cohort study. The Lancet Diabetes & Endocrinology 3, 431–436. https://doi.org/10.1016/S2213-8587(15)00033-9. Rahmouni, K., 2014. Obesity-associated hypertension: recent progress in deciphering the pathogenesis. Hypertension 64, 215–221. https://doi.org/10.1161/hypertensionaha. 114.00920. Raji, C.A., Ho, A.J., Parikshak, N.N., Becker, J.T., Lopez, O.L., Kuller, L.H., Hua, X., Leow, A.D., Toga, A.W., Thompson, P.M., 2010. Brain structure and obesity. Hum. Brain Mapp. 31, 353–364. https://doi.org/10.1002/hbm.20870. R Core Team, 2016. R: a Language and Environment for Statistical Computing. Revelle, W., 2017. Package “psych” [WWW Document]. URL. https://cran.r-project.org/ web/packages/psych/psych.pdf. Ryan, L., Walther, K., 2014. White matter integrity in older females is altered by increased body fat. Obesity 22, 2039–2046. https://doi.org/10.1002/oby.20815. Ryan, L., Walther, K., Bendlin, B.B., Lue, L.-F., Walker, D.G., Glisky, E.L., 2011. Agerelated differences in white matter integrity and cognitive function are related to APOE status. Neuroimage 54, 1565–1577. https://doi.org/10.1016/j.neuroimage. 2010.08.052. Salami, A., Eriksson, J., Nilsson, L.-G., Nyberg, L., 2012. Age-related white matter microstructural differences partly mediate age-related decline in processing speed but not cognition. Biochim. Biophys. Acta (BBA) - Mol. Basis Dis. 1822, 408–415. https:// doi.org/10.1016/j.bbadis.2011.09.001. Salthouse, T.A., 2011. What cognitive abilities are involved in trail-making performance? Intelligence 39, 222–232. https://doi.org/10.1016/j.intell.2011.03.001. Scholz, J., Klein, M.C., Behrens, T.E.J., Johansen-Berg, H., 2009. Training induces changes in white-matter architecture. Nat. Neurosci. 12, 1370–1371. https://doi.org/ 10.1038/nn.2412. Shefer, G., Marcus, Y., Stern, N., 2013. Is obesity a brain disease? Neurosci. Biobehav. Rev. 37, 2489–2503. https://doi.org/10.1016/j.neubiorev.2013.07.015.

249