Decreased Glucose Metabolism in Medial Prefrontal Areas is Associated with Nutritional Status in Patients with Prodromal and Early Alzheimer’s Disease

Abstract

Background:

Weight loss is frequently observed in patients with Alzheimer’s disease (AD); however, the underlying mechanisms are not well understood.

Objects:

To clarify the associations between nutritional status and AD-related brain changes using Pittsburgh Compound-B (PiB)-PET, fluorodeoxyglucose (FDG)-PET, and structural MRI.

Methods:

The subjects were 34 amyloid-β (Aβ)-positive individuals with mild cognitive impairment or early AD (prodromal/early AD), and 55 Aβ-negative cognitively normal (CN) subjects who attended the Multimodal Neuroimaging for AD Diagnosis (MULNIAD) study. Nutritional status of the subjects was assessed by body mass index and waist to height ratio (waist circumference/height). The associations between nutritional status and brain changes were examined by multiple regression analysis using statistical parametric mapping.

Results:

In the prodromal/early AD group, nutritional status was significantly positively correlated with regional cerebral glucose metabolism (rCGM) in the medial prefrontal cortices, while different topographical associations were seen in the CN group, suggesting these changes were AD-specific. Aβ deposition and gray matter volume were not significantly associated with nutritional status. Sub-analysis in the prodromal/early AD group demonstrated that fat mass index, but not fat-free mass index, was positively correlated with rCGM in the medial prefrontal areas.

Conclusion:

This present study provides preliminary results suggesting that hypometabolism in the medial prefrontal areas is specifically associated with AD-related weight loss, and decrease in fat mass may have a key role.

Keywords

Alzheimer’s disease composition Pittsburgh Compound B magnetic resonance imaging nutritional status positron-emission tomography

INTRODUCTION

Nutritional problems, especially weight loss (WL), are commonly seen in individuals with Alzheimer’s disease (AD). Although WL in AD is associated with adverse outcomes such as a higher rate of institutionalization and increased mortality [1, 2], the detailed mechanisms of WL remain unclear. To prevent these negative outcomes, it is important to elucidate the pathophysiological background of AD-related WL.

Many studies have demonstrated that patients with AD lose weight from the early stage, and especially in the moderate to severe stages of AD [3 –5]. In patients with AD, WL appears to be the consequence of decreased food intake due to anorexia and forgetting to eat, or exaggerated physical activity due to wandering, both of which are secondarily caused by cognitive decline or brain atrophy [6, 7]. Grundman et al. reported that atrophy of the mesial temporal lobe was associated with lower body mass index (BMI) in AD, using magnetic resonance imaging (MRI) morphometric analysis [8]. Further, ¹⁸F-fluorodeoxyglucose (FDG)-positron emission tomography (PET) results indicated that lower regional cerebral glucose metabolism (rCGM) was also associated with lower BMI in AD [9]. Moreover, another study showed that lower cerebrospinal fluid (CSF) amyloid-β (Aβ) and higher total CSF tau were associated with lower BMI in subjects with normal cognition or mild cognitive impairment (MCI) [10]. Based on these studies, the causes of WL in AD seem to be multifactorial and different in various stages of dementia, and AD-related brain changes including Aβ deposition, brain atrophy, and hypofunction influence WL.

To our knowledge, no studies have examined the association between AD-related brain changes and nutritional status comprehensively using multi imaging modalities including Aβ-PET, FDG-PET, and structural MRI. In the current study, we focused on prodromal and early-stage AD, and aimed to clarify the association between AD-related brain changes and nutritional status.

METHODS

Participants

The subjects were selected from participants in the Multimodal Neuroimaging for AD Diagnosis (MULNIAD) study, which is a prospective study implemented at the National Center for Geriatrics and Gerontology (NCGG). All participants underwent comprehensive neuropsychological batteries and neuroimaging assessments, including Pittsburgh Compound-B (PiB)-PET, FDG-PET, and structural MRI. The study was approved by the Ethics Committee of NCGG, and all participants provided written informed consent. Based on the PiB-PET imaging findings, we selected 14 Aβ-positive early AD, 20 Aβ-positive MCI (prodromal AD), and 55 Aβ-negative cognitively normal (CN) healthy elderly subjects. In this study, we analyzed prodromal AD and early AD as one group (prodromal/early AD) for the following reasons: 1) within the AD continuum of preclinical, prodromal, and dementia stages of AD, nutritional problems including appetite change are reported to start from the prodromal stage [11], and the progression of prodromal AD to the dementia stage is seamless; and 2) to increase the statistical power. Individuals under treatment for any significant medical, neurologic, or psychiatric disease, as well as those with any history of a major psychiatric disorder, were excluded.

Clinical assessment

Cognitive function was assessed by Mini-Mental State Examination (MMSE) [12], Clinical Dementia Rating (CDR) [13], and Logical Memory II score from the Wechsler Memory Scale–Revised (LM2) [14]. Based on these test results, clinical categories of early AD, prodromal AD, and CN were determined according to the inclusion criteria of the Alzheimer’s Disease Neuroimaging Initiative 2 (ADNI2) study (http://adni.loni.usc.edu/). The early AD and prodromal AD individuals also fulfilled the diagnostic criteria developed by the National Institute on Aging-Alzheimer’s Association [15, 16]. Mood was assessed using the 15-item Geriatric Depression Scale (GDS-15) [17].

Nutritional status

Nutritional status of the subjects was assessed by BMI and waist to height ratio (WHtR: [waist circumference/height]), since WHtR is known to be useful to assess abdominal fat in different age and sex groups [18]. Additionally, fat mass and fat-free mass were measured by bioelectrical impedance analysis (MC-180; Tanita Corp., Tokyo, Japan) in 19 of 34 prodromal/early AD subjects. After measurement, fat mass index (FMI) and fat-free mass index (FFMI) were calculated as fat mass and fat-free mass divided by height squared (kg/m²), respectively.

Imaging data

PET images were acquired using a PET-CT camera, Biograph True V (Siemens Healthcare, Erlangen, Germany). X-ray CT for attenuation correction was performed before PET imaging. For PiB-PET, 3D PET imaging for 50–70 min after intravenous injection of 555±185MBq [C11]-PiB was carried out. PiB-positivity was judged based on visual reading by two experienced nuclear medicine physicians, as previously described [19]. For statistical image analysis, PiB-PET images were spatially normalized with parameters obtained from individual 3D-T1 MR images co-registered to PiB-PET images using DARTEL [20]. Then, standardized uptake value ratio images were generated by dividing PiB images by the average value in both cerebellar hemispheres as previously described [19]. For FDG-PET, all participants fasted for at least 4 h, and emission was performed in the 3D mode for 30–60 min after intravenous administration of [F18]-FDG (185±37 MBq). FDG-PET images were also spatially normalized in the same manner as for PiB-PET images. High-resolution 3D T1-weighted images were acquired by MPRAGE (Magnetization-Prepared Rapid Gradient-Echo Imaging) sequence (TE/TI/TR = 2.51/900/1900 ms, 0.977×0.977×1.1 mm³) using a Siemens Trio 3T scanner. For volumetric analysis, spatially normalized gray and white matter images were created using DARTEL.

Statistical analysis

Differences in demographics between prodromal/early AD and CN were analyzed using Student’s t-test, Mann–Whitney U test or chi-squared test. With spatially normalized and smoothed PiB-PET, FDG-PET, and MR images, the relationships between nutritional status and Aβ deposition, rCGM, and gray matter volume were examined by multiple regression analysis using Statistical Parametric Mapping (SPM8; Wellcome Trust Centre for Neuroimaging., London, UK). In addition, region of interest (ROI)-based analyses for FDG-PET data were also carried out using MarsBaR toolbox (http://marsbar.sourceforge.net/). ROIs were created from thresholded statistical maps of the results of regression analyses. Then, the mean values of normalized rCGM within each ROI were extracted and used for ROI-based analyses. These analyses can be a kind of circulatory statistics [21], but were only aimed at estimating the strength of linear relationships between nutritional status and rCGM by computing Pearson correlation coefficients (r) and partial correlation coefficients (r_p) adjusted for age, sex, and education, which are not obtainable as the standard output of SPM computation.

RESULTS

Demographics of the participants are shown in Table 1. There were no differences in sex, education, and nutritional status. Prodromal/early AD patients were older (74.1±6.1 versus 69.6±5.7 y, p < 0.001), had higher GDS score (2.8±1.6 versus 1.9±1.6, p = 0.010), and had lower cognitive function (p < 0.001). APOE ɛ4 carriers comprised 79.4% of prodromal/early AD patients and 12.7% of CN subjects (p < 0.001).

Table 1

Clinical characteristics of subjects with various cognitive stages

	ALL n = 89	Prodromal/	CN n = 55	p-value
		early AD n = 34
Female, n (%)	54 (60.7)	21 (61.8)	33 (60.0)	0.868
Age, y	71.3±6.2	74.1±6.1	69.6±5.7	<0.001
Education, y	11.9±2.5	11.3±2.5	12.3±2.4	0.067
Height, cm	156.5±8.4	154.9±9.0	157.4±8.0	0.180
Weight, kg	55.3±10.4	52.6±9.9	57.0±10.4	0.055
Waist circumference, cm	83.4±8.3	82.0±7.8	84.3±8.5	0.202
BMI, kg/m²	22.5±3.1	21.8±2.8	22.9±3.2	0.108
WHtR	0.53±0.1	0.53±0.1	0.54±0.1	0.644
GDS-15, score	2.2±1.6	2.8±1.6	1.9±1.6	0.010
CDR	0.3±0.5	0.8±0.5	0.0±0.0	<0.001
MMSE, score	26.9±3.4	24.0±3.6	28.7±1.4	<0.001
Logical memory I, score	15.5±8.1	7.8±5.3	20.2±5.6	<0.001
Logical memory II, score	11.2±8.5	2.7±4.3	16.4±5.7	<0.001
ADAS-J cog, score	8.7±5.1	13.1±5.2	5.9±2.6	<0.001
APOE ɛ4 carrier, n (%)	34 (38.2)	27 (79.4)	7 (12.7)	<0.001

p values are for differences between prodromal/early AD group and CN group. CN, cognitively normal; AD, Alzheimer’s disease; BMI, body mass index; WHtR, waist to height ratio; GDS, geriatric depression scale; MMSE, Mini-Mental State Examination; CDR, clinical dementia rating; ADAS-J cog, Alzheimer’s Disease Scale for Japanese-cognitive subscale; APOE ɛ4, positive for apolipoprotein E ɛ4.

Regression analysis of PiB-PET images did not show any significant association between nutritional status and Aβ deposition in both prodromal/early AD and CN (Supplementary Figure 1A). MRI volumetric analyses also showed that gray matter volume was not significantly associated with nutritional status in both prodromal/early AD and CN (Supplementary Figure 1B). On the other hand, FDG-PET analyses demonstrated that in prodromal/early AD, BMI, and WHtR were significantly positively correlated with rCGM in the medial prefrontal areas including anterior cingulate cortex (ACC), medial prefrontal cortex (MPFC), and orbitofrontal cortex (OFC) after adjusting for age, sex, and education, whereas the CN group showed positive correlations of BMI and WHtR with rCGM in more widespread brain areas, including the ACC, middle part of the cingulate cortices, orbitofrontal cortices, and cerebellum (Fig. 1A, B). Color overlays demonstrated the differences in these regions between the prodromal/early AD and CN groups. To estimate the strength of linear relationships between nutritional status and rCGM, we performed ROI analyses. Figure 2A shows the association between BMI and rCGM. In the medial prefrontal ROI (ROI-a1), the prodromal/early AD group showed a moderately strong linear relationship between BMI and rCGM with r = 0.479 (unadjusted Pearson correlation coefficient) and r_p = 0.660 (partial correlation coefficient adjusted for age, sex, and education), whereas the CN group showed a weaker relationship (r = 0.251 and r_p = 0.261). In contrast, the CN group showed a moderately strong linear relationship (r = 0.677, r_p = 0.705) in the anterior/middle cingulate ROI (ROI-a2), whereas the prodromal/early AD group showed a weaker relationship (r = 0.296, r_p = 0279). Similar findings were also seen regarding the association between WHtR and rCGM (Fig. 2B). Moderately strong linear relationships were found in the medial prefrontal ROI (ROI-b1) for the prodromal/early AD group (r = 0.533, r_p = 0.698), and in the anterior/middle cingulate ROI (ROI-b2) for the CN group (r = 0.552, r_p = 0.702).

Fig.1

Voxel-wise multiple regression analysis of FDG-PET and nutritional status. A) Overlay of significant positive correlation of rCGM with BMI in prodromal/early AD (n = 34) and CN (n = 55) adjusted for age, sex, and education. Statistical threshold was set to FWE-corrected p < 0.05 at the height threshold of p = 0.001. Color overlays show comparison of correlation of rCGM with BMI between prodromal/early AD (red color overlay) and CN (green color overlay) (t value >3.3). B) Overlay of significant positive correlation of rCGM with WHtR in prodromal/early AD (n = 34) and CN (n = 55) adjusted for age, sex, and education. Statistical threshold was set to FWE-corrected p < 0.05 at the height threshold of p = 0.001. Color overlays show comparison of correlation of rCGM with WHtR between prodromal/early AD (red color overlay) and CN (green color overlay) (t value >3.3). C) Overlay of significant positive correlation of rCGM with body FMI (n = 19) or FFMI in 19 prodromal/early AD patients. Statistical threshold was set to FWE-corrected p < 0.05 at the height threshold of p = 0.001. BMI, body mass index; WHtR, waist to height ratio; FMI, fat mass index; FFMI, fat-free mass index; AD, Alzheimer’s disease; CN, cognitively normal; rCGM, regional cerebral glucose metabolism.

Fig.2

Results of ROI-based analyses. A) Relationships between BMI and rCGM. ROI-a1 and ROI-a2 were created based on the results of regression analyses corresponding to Fig. 1A left and Fig. 1A middle, respectively. A cluster in the medial prefrontal area (ROI-a1) or anterior/middle cingulate area (ROI-a2) was extracted from each thresholded statistical map. To create these two ROIs of approximately the same size, different thresholds were applied to the thresholded maps (p = 0.001 for ROI-a1, p = 0.000005 for ROI-a2). Scatter plots show the relationships between BMI and mean values of normalized rCGM within each ROI for the prodromal/early AD (left) and CN (right) groups. Closed circles and open triangles indicate early AD and prodromal AD individuals, respectively. Values for r and r_p show Pearson correlation coefficients (unadjusted) and partial correlation coefficients adjusted for age, sex and education, respectively. B) Relationships between WHtR and rCGM. ROI-b1 and ROI-b2 were created based on the results of regression analyses corresponding to Fig. 1B left and Fig. 1B middle, respectively. ROIs were created in the same manner as for ROI-a1 and ROI-a2, using the corresponding thresholded maps (p = 0.001 for ROI-b1, p = 0.00002 for ROI-b2). Scatter plots show the relationships between WHtR and mean values of normalized rCGM within each ROI for the prodromal/early AD (left) and CN (right) groups. ROI, region of interest; BMI, body mass index; WHtR, waist to height ratio; AD, Alzheimer’s disease; CN, cognitively normal.

Further, we conducted sub-analysis that examined the association between body composition parameters and rCGM in 19 prodromal/early AD patients. The results demonstrated that FMI, but not FFMI, was significantly positively correlated with rCGM in the MPFC (Fig. 1C).

DISCUSSION

The present study investigated the association between nutritional status and rCGM, Aβ deposition, and gray matter volume in subjects with prodromal/early AD. We found that the individual nutritional status in prodromal/early AD was positively correlated with rCGM in the medial prefrontal areas, including ACC, MPFC, and OFC, and these areas were different from those observed in CN. These results suggested that hypometabolism in the medial prefrontal areas is specifically associated with malnutrition in AD. On the other hand, we did not find any significant association between nutritional status and Aβ deposition or gray matter volume, suggesting functional but not structural alterations. To our knowledge, only one FDG-PET study has investigated the association between nutritional status and rCGM in AD [9]. The authors reported that lower BMI in AD patients was associated with hypometabolism in the ACC. Our results were consistent with this study, and strengthen the findings, because we implemented biomarker information for Aβ pathology that ensures reliable diagnosis of prodromal/early AD. Further, this study was able to investigate the association between nutritional status and AD-related brain changes in detail by utilizing multiple imaging modalities.

WL in patients with AD is considered to be related to apathy and loss of appetite [22], which are two of the most common neuropsychiatric symptoms of dementia across all stages of AD. A recent FDG-PET study demonstrated that hypometabolism in the medial prefrontal area is correlated with apathy in AD [23]. Also, Ismail et al. reported that hypoperfusion in the ACC and OFC were associated with appetite loss in AD [24]. These reports suggest that the positive correlation between rCGM and nutritional status found in our study may be related to apathy and/or loss of appetite, although we did not have information about the status of apathy and eating behavior in the subjects of this study. It is known that a depressive state is associated with reduced rCGM in the medial prefrontal areas [25]. Since the prodromal/early AD group showed significantly higher GDS scores than the CN group, the effects of mood on our findings need to be elucidated. First, we performed correlation analyses between GDS and nutritional status in prodromal/early AD. The results demonstrated that GDS was not associated with BMI (r = –0.043, p = 0.807) or with WHtR (r = –0.004, p = 0.980). Then, we conducted additional regression analyses between rCGM and nutritional status using age and GDS score as confounding covariates. The results demonstrated that the significant link between nutritional status and rCGM in the medial prefrontal areas was preserved in analyses adjusted for GDS score, indicating that GDS score does not affect the main findings (Supplementary Figure 2A).

Sub-analysis in prodromal/early AD, which was aimed to elucidate the association between body composition parameters and rCGM, demonstrated that FMI but not FFMI was positively correlated with rCGM in MPFC. This result suggests that WL in prodromal/early AD may mainly be caused by a decrease in fat mass. It has been reported that patients with AD lose waist circumference, which reflects abdominal adiposity, compared with controls [26]. Conversely, elderly individuals in the normal population mainly lose fat-free mass with aging, whereas fat mass and waist circumference rather increase [27, 28]. These observations suggest that a decrease of fat mass might be a characteristic of AD-related WL. However, it remains unclear whether this characteristic is an AD-specific change or not. In fact, Wirth et al. reported that fat mass primarily decreased with cognitive deterioration in patients with dementia in whom the clinical types were not described [29]. Therefore, we conducted additional regression analyses between rCGM and nutritional status, adjusting for cognitive decline, using MMSE score as a confounding covariate. Interestingly, the results demonstrated that WHtR still showed a significant correlation with rCGM in the MPFC, while the significant link between BMI and rCGM in the MPFC was lost by adjusting for MMSE score (Supplementary Figure 2A). Also, FMI was significantly correlated with rCGM in the MPFC after adjusting for MMSE (Supplementary Figure 2B). These results suggest that the link between abdominal adiposity and rCGM in the MPFC cannot be explained only by cognitive decline. Therefore, there might be some AD-specific mechanism. Recently, Khemka et al. reported that the serum level of leptin, which is primarily secreted by adipose tissue, is decreased in patients with AD compared with controls [30]. Together with these previous reports, our results suggest that a decrease of fat mass may have a key role in AD-related WL. Further studies to investigate the relevance of fat tissue metabolism including adipokines, brain imaging, and nutritional status may deepen our understanding of AD-related WL.

In the present study, CN subjects showed a significant positive correlation between nutritional status and rCGM in more widespread brain areas, including the anterior and middle part of the cingulate cortices, orbitofrontal cortices, and cerebellum. There are only a few studies investigating nutritional status and rCGM or cerebral blood flow (rCBF) in healthy adults. These studies reported that elevated BMI was associated with decreased rCGM/rCBF in the prefrontal cortex [31, 32]. Their results were different from our findings; however, this might be due to the difference in target population, because these studies rather focused on individuals with obesity. Using MRI volumetric analyses, Flöel et al. reported that physical activity, which is important to maintain muscle mass and to prevent sarcopenia, was associated with increased gray matter volume in the prefrontal, cingulate, occipito-temporal, and cerebellum in healthy older individuals [33]. There might be common mechanisms between their findings and our results.

Our study has several limitations. First, the sample size was relatively small. Second, because of the cross-sectional design of this study, the cause-effect relationship between nutritional status and rCGM is unclear. Given these limitations, this study revealed the association between AD-related brain changes and nutritional status comprehensively using multi imaging modalities. In conclusion, the present study suggests that hypometabolism in the medial prefrontal area is specific to nutritional problems in AD, and decrease in fat mass may have a key role.

Footnotes

ACKNOWLEDGMENTS

The authors thank all the clinical doctors, researchers, and staff at the NCGG who supported the MULNIAD project, and BioBank for quality control of the clinical data. This work was supported by grants from the Research Funding of Longevity Sciences (23–26, 25-24, 26–30, and 28–32) from the NCGG. The funding sources had no role in the study design, data collection, data analyses, or data interpretation. The MULNIAD project is registered as UMIN ID: 000006419 and 000016144.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Faxén-Irving

, Basun

, Cederholm

(2005) Nutritional and cognitive relationships and long-term mortality in patients with various dementia disorders. Age Ageing 34, 136–141.

Andrieu

, Reynish

, Nourhashemi

, Ousset

, Grandjean

, Grand

, Albarede

, Vellas

(2001) Nutritional risk factors for institutional placement in Alzheimer’s disease after one year follow-up. J Nutr Health Aging 5, 113–117.

Berkhout

, Cools

, van Houwelingen

(1998) The relationship between difficulties in feeding oneself and loss of weight in nursing-home patients with dementia. Age Ageing 27, 637–641.

Johnson

, Wilkins

, Morris

(2006) Accelerated weight loss may precede diagnosis in Alzheimer disease. Arch Neurol 63, 1312–1317.

Stewart

, Masaki

, Xue

, Peila

, Petrovitch

, White

, Launer

(2005) A 32-year prospective study of change in body weight and incident dementia: The Honolulu-Asia Aging Study. Arch Neurol 62, 55–60.

Sergi

, De Rui

, Coin

, Inelmen

, Manzato

(2013) Weight loss and Alzheimer’s disease: Temporal and aetiologic connections. Proc Nutr Soc 72, 160–165.

Droogsma

, van Asselt

, De Deyn

(2015) Weight loss andundernutrition in community-dwelling patients with Alzheimer’sdementia: From population based studies to clinical management. Z Gerontol Geriatr 48, 318–324.

Grundman

, Corey-Bloom

, Jernigan

, Archibald

, Thal

(1996) Low body weight in Alzheimer’s disease is associated with mesial temporal cortex atrophy. Neurology 46, 1585–1591.

, Okamura

, Arai

, Higuchi

, Maruyama

, Itoh

, Yamaguchi

, Sasaki

(2002) Neuroanatomical correlates of low body weight in Alzheimer’s disease: A PET study. Prog Neuropsychopharmacol Biol Psychiatry 26, 1285–1289.

10.

Ewers

, Schmitz

, Hansson

, Walsh

, Fitzpatrick

, Bennett

, Minthon

, Trojanowski

, Shaw

, Faluyi

, Vellas

, Dubois

, Blennow

, Buerger

, Teipel

, Weiner

, Hampel

, Alzheimer’s Disease Neuroimaging Initiative (2012) Body mass index is associated with biological CSF markers of core brain pathology of Alzheimer’s disease. Neurobiol Aging 33, 1599–1608.

11.

Kai

, Hashimoto

, Amano

, Tanaka

, Fukuhara

, Ikeda

(2015) Relationship between eating disturbance and dementia severity in patients with Alzheimer’s disease. PLoS One 10, e0133666.

12.

Folstein

, Folstein

, McHugh

(1975) “Mini-Mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

13.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2414.

14.

Wechsler

(1981) Wechsler Memory Scale-Revised. Psychological Corporation, San Antonio, TX.

15.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

16.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

17.

Yesavage

, Brink

, Rose

, Lum

, Huang

, Adey

, Leirer

(1982) Development and validation of a geriatric depressionscreening scale: A preliminary report. J PsychiatrRes 17, 37–49.

18.

Browning

, Hsieh

, Ashwell

(2010) A systematic review of waist-to-height ratio as a screening tool for the prediction of cardiovascular disease and diabetes: 0.5 could be a suitable global boundary value. Nutr Res Rev 23, 247–269.

19.

Kaneko

, Nakamura

, Washimi

, Kato

, Sakurai

, Arahata

, Bundo

, Takeda

, Niida

, Ito

, Toba

, Tanaka

, Yanagisawa

(2014) Novel plasma biomarker surrogating cerebral amyloid deposition. Proc Jpn Acad Ser B Phys Biol Sci 90, 353–364.

20.

Ashburner

(2007) A fast diffeomorphic image registration algorithm. Neuroimage 38, 95–113.

21.

Kriegeskorte

, Simmons

, Bellgowan

, Baker

(2009) Circular analysis in systems neuroscience: The dangers of double dipping. Nat Neurosci 12, 535–540.

22.

Volicer

, Frijters

, van der Steen

(2013) Apathy and weight loss in nursing home residents: Longitudinal study. J Am Med Dir Assoc 14, 417–420.

23.

Marshall

, Monserratt

, Harwood

, Mandelkern

, Cummings

, Sultzer

(2007) Positron emission tomography metabolic correlates of apathy in Alzheimer disease. Arch Neurol 64, 1015–1020.

24.

Ismail

, Herrmann

, Rothenburg

, Cotter

, Leibovitch

, Rafi-Tari

, Black

, Lanctôt

(2008) A functional neuroimaging study of appetite loss in Alzheimer’s disease. J Neurol Sci 271, 97–103.

25.

Boublay

, Schott

, Krolak-Salmon

(2016) Neuroimaging correlates of neuropsychiatric symptoms in Alzheimer’s disease: A review of 20 years of research. Eur J Neurol 23, 1500–1509.

26.

Chu

, Tam

, Lee

, Yik

, Song

, Cheung

, Lam

(2009) Late-life body mass index and waist circumference in amnestic mild cognitive impairment and Alzheimer’s disease. J Alzheimers Dis 17, 223–232.

27.

Yamada

, Moriguch

, Mitani

, Aoyama

, Arai

(2014) Age-dependent changes in skeletal muscle mass and visceral fat area in Japanese adults from 40 to 79 years-of-age. Geriatr Gerontol Int 14, 8–14.

28.

Kuk

, Saunders

, Davidson

, Ross

(2009) Age-related changes in total and regional fat distribution. Ageing Res Rev 8, 339–348.

29.

Wirth

, Smoliner

, Sieber

, Volkert

(2011) Cognitive function is associated with body composition and nutritional risk of geriatric patients. J Nutr Health Aging 15, 706–710.

30.

Khemka

, Bagchi

, Bandyopadhyay

, Bir

, Chattopadhyay

, Biswas

, Basu

, Chakrabarti

(2014) Altered serum levels of adipokines and insulin in probable Alzheimer’s disease. J Alzheimers Dis 41, 525–533.

31.

Volkow

, Wang

, Telang

, Fowler

, Goldstein

, Alia-Klein

, Logan

, Wong

, Thanos

, Ma

, Pradhan

(2009) Inverse association between BMI and prefrontal metabolic activity in healthy adults. Obesity (Silver Spring) 17, 60–65.

32.

Willeumier

, Taylor

, Amen

(2011) Elevated BMI is associated with decreased blood flow in the prefrontal cortex using SPECT imaging in healthy adults. Obesity (Silver Spring) 19, 1095–1097.

33.

Flöel

, Ruscheweyh

, Krüger

, Willemer

, Winter

, Völker

, Lohmann

, Zitzmann

, Mooren

, Breitenstein

, Knecht

(2010) Physical activity and memory functions: Are neurotrophins and cerebral gray matter volume the missing link? Neuroimage 49, 2756–2763.