Data_Sheet_1.PDF

Background: Changes in the metabolic and structural brain networks in mild cognitive impairment (MCI) have been widely researched. However, few studies have compared the differences in the topological properties of the metabolic and structural brain networks in patients with MCI.Methods: We analyzedmagnetic resonance imaging (MRI) and fluoro-deoxyglucose positron emission tomography (FDG-PET) data of 137 patients with MCI and 80 healthy controls (HCs). The HC group data comes from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. The permutation test was used to compare the network parameters (characteristic path length, clustering coefficient, local efficiency, and global efficiency) between the two groups. Partial Pearson’s correlation analysis was used to calculate the correlations of the changes in gray matter volume and glucose intake in the key brain regions in MCI with the Alzheimer’s Disease Assessment Scale-Cognitive (ADAS-cog) sub-item scores.Results: Significant changes in the brain network parameters (longer characteristic path length, larger clustering coefficient, and lower local efficiency and global efficiency) were greater in the structural network than in the metabolic network (longer characteristic path length) in MCI patients than in HCs. We obtained the key brain regions (left globus pallidus, right calcarine fissure and its surrounding cortex, left lingual gyrus) by scanning the hubs. The volume of gray matter atrophy in the left globus pallidus was significantly positively correlated with comprehension of spoken language (p = 0.024) and word-finding difficulty in spontaneous speech item scores (p = 0.007) in the ADAS-cog. Glucose intake in the three key brain regions was significantly negatively correlated with remembering test instructions items in ADAS-cog (p = 0.020, p = 0.014, and p = 0.008, respectively).Conclusion: Structural brain networks showed more changes than metabolic brain networks in patients with MCI. Some brain regions with significant changes in betweenness centrality in both structural and metabolic networks were associated with MCI.

Decision letter: Stage-dependent differential influence of metabolic and structural networks on memory across Alzheimer’s disease continuum

This study attempted to better understand the properties associated with the metabolic brain network in mild cognitive impairment (MCI) and Alzheimer's disease (AD). Graph theory was employed to investigate the topological organization of metabolic brain network among 86 patients with MCI, 89 patients with AD, and 97 normal controls (NCs) using 18F fluoro-deoxy-glucose positron emission tomography (FDG-PET) data. The whole brain was divided into 82 areas by Brodmann atlas to construct networks. We found that MCI and AD showed a loss of small-world properties and topological aberrations, and MCI showed an intermediate measurement between NC and AD. The networks of MCI and AD were vulnerable to attacks resulting from the altered topological pattern. Furthermore, individual contributions were correlated with Mini-Mental State Examination and Clinical Dementia Rating. The present study indicated that the topological patterns of the metabolic networks were aberrant in patients with MCI and AD, which may be particularly helpful in uncovering the pathophysiology underlying the cognitive dysfunction in MCI and AD.

The study aimed to investigate the consistency and diversity between metabolic and structural brain networks at individual level constructed with divergence-based method in healthy Chinese population. The 18F-FDG PET and T1-weighted images of brain were collected from 209 healthy participants. The Jensen-Shannon divergence (JSD) was used to calculate metabolic or structural connectivities between any pair of brain regions and then individual brain networks were constructed. The global and regional topological properties of both networks were analyzed with graph theoretical analysis. Regional properties including nodal efficiency, degree, and betweenness centrality were used to define hub regions of networks. Cross-modality similarity of brain connectivity was analyzed with differential power (DP) analysis. The default mode network (DMN) had the largest number of brain connectivities with high DP values. The small-worldness indexes of metabolic and structural networks in all participants were greater than 1. The structural network showed higher assortativity and local efficiency than metabolic network, while hierarchy and global efficiency were greater in the metabolic network (all P < 0.001). Most of hubs in both networks were symmetrically spatial distributed in the regions of the DMN and subcortical nuclei including thalamus and amygdala, etc. The human brain presented small-world architecture both in perspective of individual metabolic and structural networks. There was a structural substrate that supported the brain to globally and efficiently integrate and process metabolic interaction across brain regions. The cross-modality cooperation or specialization in both networks might imply mechanisms of achieving higher-order brain functions.

Background: Modules in brain network represent groups of brain regions that are collectively involved in one or more cognitive domains. Exploring aging-related reorganization of the brain modular architecture using metabolic brain network could further our understanding about aging-related neuromechanism and neurodegenerations. Materials and Methods: In this study, 432 subjects who performed 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET) were enrolled and divided into young and old adult groups, as well as female and male groups. The modular architecture was detected, and the connector and hub nodes were identified to explore the topological role of the brain regions based on the metabolic brain network. Results: This study revealed that human metabolic brain network was modular and could be clustered into three modules. The modular architecture was reorganized from young to old ages with regions related to sensorimotor function clustered into the same module; and the number of connector nodes was reduced and most connector nodes were localized in temporo-occipital areas related to visual and auditory functions in old ages. The major gender difference is that the metabolic brain network was delineated into four modules in old female group with the nodes related to sensorimotor function split into two modules. Discussion: Those findings suggest aging is associated with reorganized brain modular architecture. Clinical Trial Registration number: ChiCTR2000036842. Impact statement Distinguishing the basic biology underlying aging from that underlying disease is critical for the prevention, diagnosis, and treatment of the aging-related brain disorders. In this study, we tried to uncover aging-related brain modular reorganization by using metabolic brain network. We found the modular architecture was slightly reorganized from young to old ages with regions related to sensorimotor function more converged. The number of connector nodes was reduced and most connector nodes were localized into the temporo-occipital regions. The major gender difference was that metabolic brain network was delineated into four modules in the old female group with the sensorimotor functions split into two modules.

Objectives: Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder and the most common form of dementia in the older people. Some types of mild cognitive impairment (MCI) are the clinical precursors of AD, while other MCI forms tend to remain stable over time and do not progress to AD. To discriminate MCI patients at risk of AD from stable MCI, we propose a novel deep-learning radiomics (DLR) model based on 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET) images and combine DLR features with clinical parameters (DLR+C) to improve diagnostic performance.Methods: 18F-fluorodeoxyglucose positron emission tomography (PET) data from the Alzheimer's disease Neuroimaging Initiative database (ADNI) were collected, including 168 patients with MCI who converted to AD within 3 years and 187 patients with MCI without conversion within 3 years. These subjects were randomly partitioned into 90 % for the training/validation group and 10 % for the independent test group. The proposed DLR approach consists of three steps: base DL model pre-training, network features extraction, and integration of DLR+C, where a convolution network serves as a feature encoder, and a support vector machine (SVM) operated as the classifier. In comparative experiments, we compared our DLR+C method with four other methods: the standard uptake value ratio (SUVR) method, Radiomics-ROI method, Clinical method, and SUVR + Clinical method. To guarantee the robustness, 10-fold cross-validation was processed 100 times.Results: Under the DLR model, our proposed DLR+C was advantageous and yielded the best classification performance in the diagnosis of conversion with the accuracy, sensitivity, and specificity of 90.62 ± 1.16, 87.50 ± 0.00, and 93.39 ± 2.19%, respectively. In contrast, the respective accuracy of the other four methods reached 68.38 ± 1.27, 73.31 ± 6.93, 81.09 ± 1.97, and 85.35 ± 0.72 %. These results suggested the DLR approach could be used successfully in the prediction of conversion to AD, and that our proposed DLR-combined clinical information was effective.Conclusions: This study showed DLR+C could provide a novel and valuable method for the computer-assisted diagnosis of conversion to AD from MCI. This DLR+C method provided a quantitative biomarker which could predict conversion to AD in MCI patients.

Over the past two decades, the development of functional imaging methods has greatly promoted our understanding on the changes of neurons following neurodegenerative disorders, such as Parkinson's disease (PD). The application of a spatial covariance analysis on 18F-FDG PET imaging has led to the identification of a distinctive disease-related metabolic pattern. This pattern has proven to be useful in clinical diagnosis, disease progression monitoring as well as assessment of the neuronal changes before and after clinical treatment. It may potentially serve as an objective biomarker on disease progression monitoring, assessment, histological and functional evaluation of related diseases. PD is one of the most common neurodegenerative disorders in the elderly. It is characterized by progressive loss of dopamine neurons in the substantia nigra pars compacta. Throughout the course of disease, the most obvious symptoms are movement-related, such as resting tremor, muscle rigidity, hypokinesia and postural instability (Worth, 2013). Currently, a definite diagnosis of PD is made by clinical evaluation with at least 2 years of follow-up (Hughes et al., 2002; Bhidayasiri and Reichmann, 2013), due to the overlap of motor symptoms between early PD and atypical parkinsonism including multiple system atrophy (MSA) and progressive supranuclear palsy (PSP). However, this classic diagnostic criterion does not benefit the early diagnosis of disease. The prognostic outcome and treatment option are substantially different between PD and atypical parkinsonism. Thus it is critical to develop biomarkers for earlier and more accurate diagnosis of PD. Generally, appropriate diagnostic biomarker for PD ought to cover several key characteristics: (i) minimal invasiveness to detect the biomarker in easily accessible body tissue or fluids, (ii) excellent sensitivity to explore the patients with PD, (iii) high specificity to prevent false-positive results in PD-free individuals, and (iv) robustness against potential affecting factors. A PD-related spatial covariance pattern (PDRP) with quantifiable expression on 18F-FDG PET imaging has been gradually detected using a spatial covariance method during the last two decades and it has been demonstrated to be the right diagnostic biomarker for PD (Eidelberg et al., 1994). PDRP has proven not only to be effective in early discrimination of PD from atypical parkinsonian disorders, but also to be able to assess the disease progression and treatment response. Thus it is considered as a multifunctional biomarker. In this review, we aim to provide an overview of the development in pattern-based biomarker for PD.

Modular architecture of metabolic brain network and its effects on the spread of perturbation impact

Fluorodeoxyglucose (FDG) positron emission tomography (PET) is useful to predict Alzheimer's disease (AD) conversion in patients with mild cognitive impairment (MCI). However, few studies have examined the extent to which FDG PET alone can predict AD conversion and compared the efficacy between visual and computer-assisted analysis directly. The current study aimed to evaluate the value of FDG PET in predicting the conversion to AD in patients with MCI and to compare the predictive values of visual reading and computer-assisted analysis. A total of 54 patients with MCI were evaluated with FDG PET and followed-up for 2years with final diagnostic evaluation. FDG PET images were evaluated by (1) traditional visual rating, (2) composite score of visual rating of the brain cortices, and (3) composite score of computer-assisted analysis. Receiver operating characteristics (ROC) curves were compared to analyze predictive values. Nineteen patients (35.2%) converted to AD from MCI. The area under the curve (AUC) of the ROC curve of the traditional visual rating, composite score of visual rating, and computer-assisted analysis were 0.67, 0.76, and 0.79, respectively. ROC curves of the composite scores of the visual rating and computer-assisted analysis were comparable (Z = 0.463, p = 0.643). Visual rating and computer-assisted analysis of FDG PET scans were analogously accurate in predicting AD conversion in patients with MCI. Therefore, FDG PET may be a useful tool for screening AD conversion in patients with MCI, when using composite score, regardless of the method of interpretation.

Background and objectives: Mild cognitive impairment (MCI) is an intermediate state between normal aging and dementia, and can be divided into amnestic and non-amnestic types. Patients with amnestic MCI present with memory impairments that are often considered as the early manifestation of Alzheimer’s disease. Patients with amnestic MCI are more likely to progress to Alzheimer’s disease than patients with non-amnestic MCI. The U.S. Food and Drug Administration has not yet approved any drug that can treat amnestic MCI. Moxibustion is a common noninvasive traditional oriental intervention, which uses mainly the heat generated by burning herbal preparations containing moxa and mugwort (Artemisia vulgaris) to simulate acupoints for alleviating the symptoms. To date, many clinical studies have investigated the clinical use of moxibustion to improve memory impairments of Alzheimer’s disease, but these have failed to make a distinction between amnestic and non-amnestic MCI. Therefore, this trial has been designed to assess the effectiveness of moxibustion on amnestic MCI using the Montreal Cognitive Assessment Scale. We will also assess the safety of moxibustion in healthy controls, and analyze the variation of brain functional connectivity and effective brain networks in patients with amnestic MCI undergoing moxibustion using function magnetic resonance imaging. Design: This is a prospective, single-center, randomized controlled clinical trial. Methods: This study will enroll 64 patients with amnestic MCI and 48 healthy controls at Baoan People’s Hospital, Shenzhen, China. The first 64 recruited patients with amnestic MCI will be randomly divided into moxibustion, placebo moxibustion, drug, and control groups (n = 16 per group). In the moxibustion group, patients will be given moxa-wool moxibustion for 12 consecutive weeks. The placebo moxibustion group will receive placebo moxibustion on the same acupoints. Patients in the drug group will be given oral administration of donepezil hydrochloride tablets, 5 mg daily, for 12 consecutive weeks. The control group will receive no intervention. Forty-eight healthy controls will also be randomly assigned into moxibustion, placebo moxibustion, and control groups (n = 16 per group). Interventions will be the same as those received by the patients with amnestic MCI. Evaluators will be blind to group allocation. Outcome measures and preliminary results: The primary outcome measure will be the improvement in cognitive function 12 months after treatment. Secondary outcome measures will be the scores on the Montreal Cognitive Assessment Scale, Clinical Dementia Rating Scale, Mini-Mental State Examination Scale, and Activity of Daily Living Scale before treatment, after 12 weeks of treatment, and 6 months after the end of treatment, as well as brain function analysis before treatment and after 12 weeks of treatment and adverse events during treatment and follow-up. A correlation analysis between cognitive function scores and brain function results will be performed. Results of our preliminary study involving 60 patients with amnestic MCI who experienced moxibustion or received no treatment showed that moxibustion on acupoints significantly improved cognitive ability and quality of sleep in patients relative to the baseline and compared with the control group. Moreover, the scores on attention and delayed recall in the moxibustion group after treatment were significantly higher than those at base line. In the control group the scores on visual space, execution, and delayed recall were significantly lower than those at baseline. These findings indicate that moxibustion improves patient’s attention and delayed recall. If not, visual space, execution, and delayed recall in MCI patients tend to be declined over time. No obvious adverse responses to moxibustion treatment occurred in the preliminary study. Discussion: This proposed trial has the potential to uncover that moxibustion will enhance cognitive-related brain function connections and effector brain networks, which is not yet known. If moxibustion is shown to be an effective and safe treatment strategy in patients with amnestic MCI, then this may pave the way for use of this treatment in clinic amnestic MCI. Ethics and dissemination: This study was approved by the Medical Ethics Committee of the Chinese Clinical Trial Registry (approval No. ChiECRCT-2017018) in October 2016, and registered on April 2017. The study was designed in June 2016. Patient recruitment was initialized in October 2016. Data analysis will be completed in December 2019. Results will be disseminated through publication in a peer-reviewed journal. Protocol version: 1.0. Trial registration: This trial was registered with the Chinese Clinical Trial Registry (registration number: ChiCTR-POC-17011162).

PurposePositron emission tomography (PET) with 18F-fluorodeoxyglucose (FDG) reveals altered cerebral metabolism in individuals with mild cognitive impairment (MCI) and Alzheimer’s dementia (AD). Previous metabolic connectome analyses derive from groups of patients but do not support the prediction of an individual’s risk of conversion from present MCI to AD. We now present an individual metabolic connectome method, namely the Kullback-Leibler Divergence Similarity Estimation (KLSE), to characterize brain-wide metabolic networks that predict an individual’s risk of conversion from MCI to AD.MethodsFDG-PET data consisting of 50 healthy controls, 332 patients with stable MCI, 178 MCI patients progressing to AD, and 50 AD patients were recruited from ADNI database. Each individual’s metabolic brain network was ascertained using the KLSE method. We compared intra- and intergroup similarity and difference between the KLSE matrix and group-level matrix, and then evaluated the network stability and inter-individual variation of KLSE. The multivariate Cox proportional hazards model and Harrell’s concordance index (C-index) were employed to assess the prediction performance of KLSE and other clinical characteristics.ResultsThe KLSE method captures more pathological connectivity in the parietal and temporal lobes relative to the typical group-level method, and yields detailed individual information, while possessing greater stability of network organization (within-group similarity coefficient, 0.789 for sMCI and 0.731 for pMCI). Metabolic connectome expression was a superior predictor of conversion than were other clinical assessments (hazard ratio (HR) = 3.55; 95% CI, 2.77–4.55; P < 0.001). The predictive performance improved further upon combining clinical variables in the Cox model, i.e., C-indices 0.728 (clinical), 0.730 (group-level pattern model), 0.750 (imaging connectome), and 0.794 (the combined model).ConclusionThe KLSE indicator identifies abnormal brain networks predicting an individual’s risk of conversion from MCI to AD, thus potentially constituting a clinically applicable imaging biomarker.

BackgroundThe default‐mode network (DMN) consists of brain regions with higher resting activity levels. Amyloid‐ß (Aß) deposition in Alzheimer’s disease (AD) occurs predominantly throughout the DMN, suggesting that activity within the network may facilitate disease processes. Indeed, increased neural activity is positively associated with Aß production. In this context, variations in DMN activity and associated metabolic networks may be linked to the risk of developing AD. However, how patterns of metabolic disruption relate to the progression of AD pathology remains unknown. Here, we investigated whether the metabolic brain networks (MBNs) architecture predicts clinical conversion in cognitively unimpaired (CU) individuals.MethodWe selected CU individuals negative to amyloid and tau (A‐T‐) from the ADNI cohort with [18F]FDG‐PET imaging data at baseline. These patients were divided in stable (non‐converters, n = 18) and clinical progressors (converters, n = 22). Individuals were age‐ and APOEe4‐matched (Table 1). The mean [18F]FDG standard uptake value ratio (SUVR, pons as reference) of brain regions of interest (ROIs) was extracted based on the DKT atlas. MBNs were assembled with a multiple sampling bootstrap scheme and corrected for group imbalance with the Adaptive Synthetic Sampling Approach for Imbalance (ADASYN) and for multiple comparisons using FDR (p &lt; 0.05).Result[18F]FDG regional SUVRs presented no differences between groups (Figure 1). However, converters had a prominent brain PET metabolic hyperconnectivity compared to non‐converters, with a 1.5 fold‐change in connection density (p &lt; 0.001, Figure 2A). Notably, this hyperactivation was not limited to the ROIs comprising the DMN; MBNs constructed with all brain regions reveal that the brains of converters typically display metabolic hyperactivity before the onset of CI (Figure 2B).ConclusionOur findings suggest the existence of early metabolic alterations at the network level in amyloid negative converters. This corroborates the notion that early soluble forms of amyloid, considered synaptoxins, may trigger brain metabolic hyperconnectivity. MBNs hold promise as biomarkers for detecting individuals at risk of clinical progression, even before amyloid positivity status.

BackgroundThe default‐mode network (DMN) consists of brain regions with higher resting activity levels. Amyloid‐β (Aβ) deposition in Alzheimer’s disease (AD) occurs predominantly throughout the DMN, suggesting that activity within the network may facilitate disease processes. Indeed, increased neural activity is positively associated with Aβ production. In this context, variations in DMN activity and associated metabolic networks may be linked to the risk of developing AD. However, how patterns of metabolic disruption relate to the progression of AD pathology remains unknown. Here, we investigated whether the metabolic brain networks (MBNs) architecture predicts clinical conversion in cognitively unimpaired (CU) individuals.MethodWe selected CU individuals negative to amyloid and tau (A‐T‐) from the ADNI cohort with [18F]FDG‐PET imaging data at baseline. These patients were divided in stable (non‐converters, n = 18) and clinical progressors (converters, n = 22). Individuals were age‐ and APOEε4‐matched (Table 1). The mean [18F]FDG standard uptake value ratio (SUVR, pons as reference) of brain regions of interest (ROIs) was extracted based on the DKT atlas. MBNs were assembled with a multiple sampling bootstrap scheme and corrected for group imbalance with the Adaptive Synthetic Sampling Approach for Imbalance (ADASYN) and for multiple comparisons using FDR (p &lt; 0.05).Result[18F]FDG regional SUVRs presented no differences between groups (Figure 1). However, converters had a prominent brain PET metabolic hyperconnectivity compared to non‐converters, with a 1.5 fold‐change in connection density (p &lt; 0.001, Figure 2A). Notably, this hyperactivation was not limited to the ROIs comprising the DMN; MBNs constructed with all brain regions reveal that the brains of converters typically display metabolic hyperactivity before the onset of CI (Figure 2B).ConclusionOur findings suggest the existence of early metabolic alterations at the network level in amyloid negative converters. This corroborates the notion that early soluble forms of amyloid, considered synaptoxins, may trigger brain metabolic hyperconnectivity. MBNs hold promise as biomarkers for detecting individuals at risk of clinical progression, even before amyloid positivity status.

BackgroundThe availability of 18‐F fluorodeoxyglucose positron emission tomography (FDG‐PET) is not universal. We hypothesized that synthetically generated FDG‐PET images would be as sensitive to detecting the pattern of hypometabolism associated with AD as real images.MethodWe developed a deep learning‐based method to produce synthetic FDG‐PET images from 1,828 T1‐weighted MRI / real FDG‐PET image pairs from the ADNI dataset, and validated the technique on a further 284 image pairs. The technique generated synthetic FDG‐PET images which were then processed to compare Standardized Uptake Value Ratio (SUVR) with the pons as reference in 81 brain regions as defined in the Desikan‐Killiany‐Tourville and subcortical default FreeSurfer atlases.ResultWe tested the differences between synthetic and real FDG‐PET on 745 image pairs (205 controls, 365 mild cognitive impairment (MCI) and 175 AD)(Table 1). Correlations in SUVR values between synthetic and real FDG‐PET ranged between weak (r = 0.13) to strong (r = 0.63), with moderate results in key regions for AD (bilateral precuneus, r = 0.43; bilateral posterior cingulate, r = 0.37). There were significant between‐group (control vs MCI and control vs AD) differences in SUVR values for all regions between synthetic and real PET‐FDG (Figure 1) with synthetic FDG‐PET having lower values. Inter‐group effect sizes were not significantly different in the majority of brain regions (76/81)(Figure 2), with similar effect sizes in the right precuneus (synthetic: ‐0.98 vs original: ‐1.37), left (‐0.77 vs ‐1.0459) and right (‐0.80 vs ‐1.04) posterior cingulate, but different for the left precuneus (‐0.91 vs ‐1.33).ConclusionSynthetic images would increase patients’ accessibility to a meaningful modality for disease assessment while decreasing their exposure to radiation and resources in the health care system.

BackgroundThe availability of 18‐F fluorodeoxyglucose positron emission tomography (FDG‐PET) is not universal. We hypothesized that synthetically generated FDG‐PET images would be as sensitive to detecting the pattern of hypometabolism associated with AD as real images.MethodWe developed a deep learning‐based method to produce synthetic FDG‐PET images from 1,828 T1‐weighted MRI / real FDG‐PET image pairs from the ADNI dataset, and validated the technique on a further 284 image pairs. The technique generated synthetic FDG‐PET images which were then processed to compare Standardized Uptake Value Ratio (SUVR) with the pons as reference in 81 brain regions as defined in the Desikan‐Killiany‐Tourville and subcortical default FreeSurfer atlases.ResultWe tested the differences between synthetic and real FDG‐PET on 745 image pairs (205 controls, 365 mild cognitive impairment (MCI) and 175 AD)(Table 1). Correlations in SUVR values between synthetic and real FDG‐PET ranged between weak (r = 0.13) to strong (r = 0.63), with moderate results in key regions for AD (bilateral precuneus, r = 0.43; bilateral posterior cingulate, r = 0.37). There were significant between‐group (control vs MCI and control vs AD) differences in SUVR values for all regions between synthetic and real PET‐FDG (Figure 1) with synthetic FDG‐PET having lower values. Inter‐group effect sizes were not significantly different in the majority of brain regions (76/81)(Figure 2), with similar effect sizes in the right precuneus (synthetic: ‐0.98 vs original: ‐1.37), left (‐0.77 vs ‐1.0459) and right (‐0.80 vs ‐1.04) posterior cingulate, but different for the left precuneus (‐0.91 vs ‐1.33).ConclusionSynthetic images would increase patients’ accessibility to a meaningful modality for disease assessment while decreasing their exposure to radiation and resources in the health care system.