Alzheimer's Disease Neuroimaging Initiative Dataset Research Articles

Light-weight cross-view hierarchical fusion network for joint localization and identification in Alzheimer’s disease with adaptive instance-declined pruning

Magnetic resonance imaging (MRI) has been widely used in assessing development of Alzheimer’s disease (AD) by providing structural information of disease-associated regions (e.g. atrophic regions). In this paper, we propose a light-weight cross-view hierarchical fusion network (CvHF-net), consisting of local patch and global subject subnets, for joint localization and identification of the discriminative local patches and regions in the whole brain MRI, upon which feature representations are then jointly learned and fused to construct hierarchical classification models for AD diagnosis. Firstly, based on the extracted class-discriminative 3D patches, we employ the local patch subnets to utilize multiple 2D views to represent 3D patches by using an attention-aware hierarchical fusion structure in a divide-and-conquer manner. Since different local patches are with various abilities in AD identification, the global subject subnet is developed to bias the allocation of available resources towards the most informative parts among these local patches to obtain global information for AD identification. Besides, an instance declined pruning algorithm is embedded in the CvHF-net for adaptively selecting most discriminant patches in a task-driven manner. The proposed method was evaluated on the AD Neuroimaging Initiative dataset and the experimental results show that our proposed method can achieve good performance on AD diagnosis.

A CNN based framework for classification of Alzheimer’s disease

In the current decade, advances in health care are attracting widespread interest due to their contributions to people longer surviving and fitter lives. Alzheimer’s disease (AD) is the commonest neurodegenerative and dementing disease. The monetary value of caring for Alzheimer’s disease patients is involved to rise dramatically. The necessity of having a computer-aided system for early and accurate AD classification becomes crucial. Deep-learning algorithms have notable advantages rather than machine learning methods. Many recent research studies that have used brain MRI scans and convolutional neural networks (CNN) achieved promising results for the diagnosis of Alzheimer’s disease. Accordingly, this study proposes a CNN based end-to-end framework for AD-classification. The proposed framework achieved 99.6%, 99.8%, and 97.8% classification accuracies on Alzheimer’s disease Neuroimaging Initiative (ADNI) dataset for the binary classification of AD and Cognitively Normal (CN). In multi-classification experiments, the proposed framework achieved 97.5% classification accuracy on the ADNI dataset.

Navigating Neuroimaging Datasets ADNI for Alzheimer's Disease

Many researchers have used the Alzheimer's Disease Neuroimaging Initiative (ADNI) to project Alzheimer's disease patients' brain images. However, the learning curve required to use the ADNI is not trivial; supplements are required to work with the complexity of the subject matter and convoluted user interface in times of insufficient software extensions or technical support. Meanwhile, papers published about the ADNI have involved popular artificial intelligence models while showing uneven performance, many supplying incomparable accuracy figures. A tutorial is thus offered, beyond what can be found in the ADNI website, for furthering research about Alzheimer's disease. It also recommends areas of enhancements to make ADNI datasets more user-friendly.

Indication of retrograde tau spreading along Braak stages and functional connectivity pathways.

Tau pathology progression in Alzheimer's disease (AD) is explained through the network degeneration hypothesis and the neuropathological Braak stages; however, the compatibility of these models remains unclear. We utilized [18F]AV-1451 tau-PET scans of 39 subjects with AD and 39 sex-matched amyloid-negative healthy controls (HC) in the ADNI (Alzheimer's Disease Neuroimaging Initiative) dataset. The peak cluster of tau-tracer uptake was identified in each Braak stage of neuropathological tau deposition and used to create a seed-based functional connectivity network (FCN) using 198 HC subjects, to identify healthy networks unaffected by neurodegeneration. Voxel-wise tau deposition was both significantly higher inside relative to outside FCNs and correlated significantly and positively with levels of healthy functional connectivity. Within many isolated Braak stages and regions, the correlation between tau and intrinsic functional connectivity was significantly stronger than it was across the whole brain. In this way, each peak cluster of tau was related to multiple Braak stages traditionally associated with both earlier and later stages of disease. We show specificity of healthy FCN topography for AD-pathological tau as well as positive voxel-by-voxel correlations between pathological tau and healthy functional connectivity. We propose a model of "up- and downstream" functional tau progression, suggesting that tau pathology evolves along functional connectivity networks not only "downstream" (i.e., along the expected sequence of the established Braak stages) but also in part "upstream" or "retrograde" (i.e., against the expected sequence of the established Braak stages), with pathology in earlier Braak stages intensified by its functional relationship to later disease stages.

Harmonizing the preclinical Alzheimer cognitive composite for multi‐cohort studies

AbstractBackgroundVersions of the Preclinical Alzheimer Cognitive Composite (PACC) and within cohort standardization models have been implemented in multiple observational and clinical trials. Harmonizing PACC data across studies, however, requires modern psychometric approaches to reflect the psychometric, testing, and demographic nuances between cohorts and enable meta‐analyses. We present a method for harmonizing PACC across four cohorts: the Alzheimer’s Disease Neuroimaging Initiative (ADNI), the Anti‐Amyloid Treatment in Asymptomatic Alzheimer’s disease (A4), Harvard Aging Brain Study (HABS), and Australian Imaging, Biomarker & Lifestyle Study of Ageing (AIBL).MethodTables 1 and 2 show the demographic and neuropsychological test data used to form each PACC. The ADNI dataset was defined as the anchor for harmonization. The final visit data were used to set factor score loadings for each PACC (except A4) to maximize cognitive variability in the sample for calibration purposes. First, we used confirmatory factor analysis (CFA) on the ADNI dataset to ascertain raw loadings and thresholds of each test onto a latent PACC factor (lPACC). These loadings were then applied to the longitudinal ADNI dataset (Fig.1). For the other studies, common items (MMSE and Logical Memory) were set as “anchors” to freely estimate unique items. Common items in HABS were set as anchors for A4 (FCsrt and DSST). Longitudinal lPACC scores were generated once all item parameters were estimated for all cohorts. We performed validation analyses to assess the quality of integration using lPACC scores versus within‐cohort standardized PACC (zPACC).ResultLPACC scores largely aligned with zPACC, particularly when examining within‐cohort relationships with baseline demographic variables (Table3) and comparing slope distributions across cohorts (Fig2B). Notable differences between the scores include: a) shifted baseline distributions for lPACC relative to zPACC, reflecting cohort demographic heterogeneity (Fig.2A), and b) constricted slope distribution for lPACC slopes compared with zPACC slopes (Fig.2B), suggesting reduced variability (Fig.2B/C). When the cohorts were combined, the point estimate for strength of association between baseline amyloid and longitudinal cognitive change was higher using the lPACC than the zPACC (Fig.3).ConclusionLPACC scores reflect within cohort patterns, while reducing slope variability. Using advanced psychometrics, a harmonized PACC can be pooled across studies, facilitating robust multi‐cohort analyses.

Validating digital cognitive biomarkers for Alzheimer’s disease

AbstractBackgroundThis study sought to validate a pragmatic method to predict impending cognitive decline in Alzheimer’s disease (AD) patients, prior to the development of hallmark symptoms. The method, a Hierarchical Bayesian Cognitive Process (HBCP) model, uses item responses to a wordlist memory (WLM) test to generate digital cognitive biomarkers (Table 1). We replicated our earlier study, performed on AVLT item response data from Mayo Clinic, by applying the HBCP model to distinguish non‐decliner and decliner groups of normal subjects at baseline from a novel dataset, the Alzheimer’s Disease Neuroimaging Initiative (ADNI), with a distinct WLM test, ADAS‐Cog.MethodFrom the ADNI dataset, we classified non‐decliner subjects (n = 442) as those whose diagnosis remained normal for 3 or more years after normal baseline assessment and decliner subjects (n = 61) as those who developed amnestic MCI or AD within 3 years of normal baseline assessment. Three analytic approaches were compared: 1) Traditional summary scores per ADAS‐Cog task were assessed for group differences with a commonly‐used statistical approach, logistic regression; 2) summary scores were assessed with a Bayesian modeling approach; and 3) item response data were assessed with our HBCP model, generating digital cognitive biomarkers.ResultLogistic regression of summary scores generated β coefficients (Table 2) that did not significantly discriminate between groups. Bayes Factor assessment of fitted Gaussian distributions to the per‐task group recall by summary score measurement provided moderate evidence that the groups were measurably the same (BF sd = 3.4, 3.1, 2.9, and 1.4, respectively). The HBCP model produced posterior distributions of group differences. Bayes Factor assessment identified three with notable group differences: Immediate Retrieval from Durable Storage, L1 (BF ds = 11.8, strong evidence), One‐shot Learning, r (BF ds = 4.5, moderate), and Partial Learning, a (BF ds = 2.9, weak; Figure).ConclusionThe obtained strong evidence for L1 group differences represents successful replication of previous work with our HBCP model on this novel dataset. The ability to identify differences in subjects diagnosed cognitively normal at time of assessment demonstrates the HBCP model’s advantage over summary score assessment and its applicability for detection of impending cognitive decline in asymptomatic AD patients.

Zero Echo Time MRAC on FDG-PET/MR Maintains Diagnostic Accuracy for Alzheimer's Disease; A Simulation Study Combining ADNI-Data.

AimAttenuation correction using zero-echo time (ZTE) – magnetic resonance imaging (MRI) (ZTE-MRAC) has become one of the standard methods for brain-positron emission tomography (PET) on commercial PET/MR scanners. Although the accuracy of the net tracer-uptake quantification based on ZTE-MRAC has been validated, that of the diagnosis for dementia has not yet been clarified, especially in terms of automated statistical analysis. The aim of this study was to clarify the impact of ZTE-MRAC on the diagnosis of Alzheimer’s disease (AD) by performing simulation study.MethodsWe recruited 27 subjects, who underwent both PET/computed tomography (CT) and PET/MR (GE SIGNA) examinations. Additionally, we extracted 107 subjects from the Alzheimer Disease Neuroimaging Initiative (ADNI) dataset. From the PET raw data acquired on PET/MR, three FDG-PET series were generated, using two vendor-provided MRAC methods (ZTE and Atlas) and CT-based AC. Following spatial normalization to Montreal Neurological Institute (MNI) space, we calculated each patient’s specific error maps, which correspond to the difference between the PET image corrected using the CTAC method and the PET images corrected using the MRAC methods. To simulate PET maps as if ADNI data had been corrected using MRAC methods, we multiplied each of these 27 error maps with each of the 107 ADNI cases in MNI space. To evaluate the probability of AD in each resulting image, we calculated a cumulative t-value using a fully automated method which had been validated not only in the original ADNI dataset but several multi-center studies. In the method, PET score = 1 is the 95% prediction limit of AD. PET score and diagnostic accuracy for the discrimination of AD were evaluated in simulated images using the original ADNI dataset as reference.ResultsPositron emission tomography score was slightly underestimated both in ZTE and Atlas group compared with reference CTAC (−0.0796 ± 0.0938 vs. −0.0784 ± 0.1724). The absolute error of PET score was lower in ZTE than Atlas group (0.098 ± 0.075 vs. 0.145 ± 0.122, p < 0.001). A higher correlation to the original PET score was observed in ZTE vs. Atlas group (R2: 0.982 vs. 0.961). The accuracy for the discrimination of AD patients from normal control was maintained in ZTE and Atlas compared to CTAC (ZTE vs. Atlas. vs. original; 82.5% vs. 82.1% vs. 83.2% (CI 81.8–84.5%), respectively).ConclusionFor FDG-PET images on PET/MR, attenuation correction using ZTE-MRI had superior accuracy to an atlas-based method in classification for dementia. ZTE maintains the diagnostic accuracy for AD.

Hi-GCN: A hierarchical graph convolution network for graph embedding learning of brain network and brain disorders prediction

PurposeRecently, brain connectivity networks have been used for the classification of neurological disorder, such as Autism Spectrum Disorders (ASD) or Alzheimer's disease (AD). Network analysis provides a new way for exploring the association between brain functional deficits and the underlying structural disruption related to brain disorders. Network embedding learning that aims to automatically learn low-dimensional representations for brain networks has drawn increasing attention in recent years. MethodIn this work we build upon graph neural network in order to learn useful representations for graph classification in an end-to-end fashion. Specifically, we propose a hierarchical GCN framework (called hi-GCN) to learn the graph feature embedding while considering the network topology information and subject's association at the same time. ResultsTo demonstrate the effectiveness of our approach, we evaluate the performance of the proposed method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and Autism Brain Imaging Data Exchange (ABIDE) dataset. Extensive experiments on ABIDE and ADNI datasets have demonstrated competitive performance of the hi-GCN model. Specifically, we obtain an average accuracy of 73.1%/78.5% as well as AUC of 82.3%/86.5% on ABIDE/ADNI. The comprehensive experiments demonstrate that our hi-GCN is effective for graph classification with brain disorders diagnosis. ConclusionThe proposed hi-GCN method performs the graph embedding learning from a hierarchical perspective while considering the structure in individual brain network and the subject's correlation in the global population network, which can capture the most essential embedding features to improve the classification performance of disease diagnosis. Moreover, the proposed jointly optimizing strategy also achieves faster training and easier convergence than both the hi-GCN with pre-training and two-step supervision.

Slice-selective learning for Alzheimer's disease classification using a generative adversarial network: a feasibility study of external validation.

The aim of this feasibility study was to use slice selective learning using a Generative Adversarial Network for external validation. We aimed to build a model less sensitive to PET imaging acquisition environment, since differences in environments negatively influence network performance. To investigate the slice performance, each slice evaluation was performed. We trained our model using a 18F-fluorodeoxyglucose ([18F]FDG) PET/CT dataset obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database and tested the model with a Severance Hospital dataset. We applied slice selective learning to reduce computational cost and to extract unbiased features. We extracted features of Alzheimer's disease (AD) and normal cognitive (NC) condition using a Boundary Equilibrium Generative Adversarial Network (BEGAN) for stable convergence. Then, we utilized these features to train a support vector machine (SVM) classifier to distinguish AD from NC. The slice range that covered the posterior cingulate cortex (PCC) using double slices showed the best performance. The accuracy, sensitivity, and specificity of our proposed network was 94.33%, 91.78%, and 97.06% using the Severance dataset and 94.82%, 92.11%, and 97.45% using the ADNI dataset. The performance on the two independent datasets showed no statistical difference (p > 0.05). Moreover, there was a statistical difference in the performance between using two slices and one slice as input (p < 0.05). Our model learned the generalized features of AD and NC for external validation when appropriate slices were selected. This study showed the feasibility of this model with consistent performance when tested using datasets acquired from a variety of image-acquisition environments.

Detecting Alzheimer's disease Based on 4D fMRI: An exploration under deep learning framework

Applying machine learning methods to various modality medical images and clinical data for early diagnosis of Alzheimer's disease (AD) and its prodromal stage has many significant results. So far, the image data input to classifier mainly focus on 2D or 3D images. Although some functional imaging technologies, such as functional magnetic resonance imaging (fMRI), generate 4D data which contain both spatial and time-varying information of the brain, for the lack of suitable 4D image processing algorithm, these 4D data were always used by transforming them into functional connectivity or slicing them into 2D/3D pictures which causing apparent information loss. In this paper, we present a 4D deep learning model (C3d-LSTM) for AD discrimination, which is able to utilize the spatial and time-varying information simultaneously by dealing with 4D fMRI data directly. The proposed C3d-LSTM combines a series of 3D convolutional neural networks (CNNs) to extract spatial features from each volume of 3D static image in fMRI image sequence. Then the feature maps obtained were put into the long short-term memory (LSTM) network to capture the time-varying information contained in the data. Because of the design of structure, C3d-LSTM became an end-to-end data-driven model, which was more convenient for processing 4D fMRI data. The model proposed conducted on the AD Neuroimaging Initiative dataset for algorithm evaluation compared with controlled experiments. Results showed that using 4D fMRI data directly with the proposed method did make a far better result for AD detection than the methods using functional connectivity, 2D, or 3D fMRI data. It demonstrated our assumption that making the most of the natural spatial and temporal information preserved in 4D fMRI data is significant for AD detection and can increase the performance of the classifier. Meanwhile the result implied that the C3d-LSTM model proposed is a suitable model for processing 4D fMRI data and extracting the spatio-temporal property of fMRI data fully for diagnosis of AD.

Choroid plexus volume is associated with levels of CSF proteins: relevance for Alzheimer's and Parkinson's disease

The choroid plexus (ChP) is a major source of cerebrospinal fluid (CSF) production, with a direct and indirect role in protein clearance, and pathogenesis of Alzheimer's disease (AD). Here, we tested the link between the ChP volume and levels of CSF proteins in 2 data sets of (i) healthy controls, mild cognitive impairment (MCI), and AD patients from the Alzheimer's Disease Neuroimaging Initiative (ADNI) (N = 509), and (ii) healthy controls and Parkinson's disease (PD) patients from the Parkinson's Progression Markers Initiative (N = 302). All patients had baseline CSF proteins (amyloid-β, total and phosphorylated-tau and α-synuclein (only in Parkinson's Progression Markers Initiative)). ChP was automatically segmented on 3T structural T1-weighted MRIs. We found negative associations between ChP volume and CSF proteins, which were stronger in healthy controls, early-MCI patients, and PD patients compared with late-MCI and AD patients. Further grouping of patients of ADNI dataset into amyloid-positive and amyloid-negative based on their florbetapir (AV45) PET imaging showed that the association between ChP volume and CSF proteins (t/p-tau) was lower in amyloid-positive group. Our findings support the possible role of ChP in the clearance of CSF proteins, provide evidence for ChP dysfunction in AD, and suggest the need to account for the ChP volume in future studies of CSF-based biomarkers.

Editors' note: Type 2 diabetes mellitus, brain atrophy, and cognitive decline

In the article “Type 2 diabetes mellitus, brain atrophy, and cognitive decline,” Moran et al. examined longitudinal relationships between type 2 diabetes mellitus (T2DM), cortical thickness, and cognitive function in older people with normal cognition, mild cognitive impairment (MCI), and Alzheimer disease (AD) in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. They concluded that T2DM contributed to cognitive decline via neurodegeneration (lower baseline cortical thickness) and that previous brain and cognitive reserve (reflected by greater education) may protect against this effect. In response, Ba et al. argue that it is unclear whether the authors properly controlled for other risk factors for cognitive decline such as hypertension, hyperlipidemia, white matter lesion burden, and depression and note that incorporating additional T2DM-related factors such as fasting blood glucose or disease duration would be relevant. They also critique the study's reliance on clinical diagnoses for determination of MCI and AD, which can often result in misattribution of symptoms to AD when supportive biomarkers are not included. Replying to these comments, Dr. Moran acknowledges the absence of detailed T2DM-specific information as a limitation of repurposing the ADNI dataset for their work but notes that the study reported data on fasting glucose, diabetes medication use, blood pressure, and measures of obesity. Although ADNI participants were selected to have a low burden of cerebrovascular disease, Dr. Moran acknowledges that there are additional potential confounders of the relationship between T2DM and brain health, such as other vascular risk factors, some of which were explored in other work using samples with a greater burden of cerebrovascular disease. He also agrees on the importance of incorporating biomarkers of underlying pathology into future work in this area. In the article “Type 2 diabetes mellitus, brain atrophy, and cognitive decline,” Moran et al. examined longitudinal relationships between type 2 diabetes mellitus (T2DM), cortical thickness, and cognitive function in older people with normal cognition, mild cognitive impairment (MCI), and Alzheimer disease (AD) in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. They concluded that T2DM contributed to cognitive decline via neurodegeneration (lower baseline cortical thickness) and that previous brain and cognitive reserve (reflected by greater education) may protect against this effect. In response, Ba et al. argue that it is unclear whether the authors properly controlled for other risk factors for cognitive decline such as hypertension, hyperlipidemia, white matter lesion burden, and depression and note that incorporating additional T2DM-related factors such as fasting blood glucose or disease duration would be relevant. They also critique the study's reliance on clinical diagnoses for determination of MCI and AD, which can often result in misattribution of symptoms to AD when supportive biomarkers are not included. Replying to these comments, Dr. Moran acknowledges the absence of detailed T2DM-specific information as a limitation of repurposing the ADNI dataset for their work but notes that the study reported data on fasting glucose, diabetes medication use, blood pressure, and measures of obesity. Although ADNI participants were selected to have a low burden of cerebrovascular disease, Dr. Moran acknowledges that there are additional potential confounders of the relationship between T2DM and brain health, such as other vascular risk factors, some of which were explored in other work using samples with a greater burden of cerebrovascular disease. He also agrees on the importance of incorporating biomarkers of underlying pathology into future work in this area.

Multi-modal neuroimaging feature selection with consistent metric constraint for diagnosis of Alzheimer's disease.

The accurate diagnosis of Alzheimer's disease (AD) and its early stage, e.g., mild cognitive impairment (MCI), is essential for timely treatment or possible intervention to slow down AD progression. Recent studies have demonstrated that multiple neuroimaging and biological measures contain complementary information for diagnosis and prognosis. Therefore, information fusion strategies with multi-modal neuroimaging data, such as voxel-based measures extracted from structural MRI (VBM-MRI) and fluorodeoxyglucose positron emission tomography (FDG-PET), have shown their effectiveness for AD diagnosis. However, most existing methods are proposed to simply integrate the multi-modal data, but do not make full use of structure information across the different modalities. In this paper, we propose a novel multi-modal neuroimaging feature selection method with consistent metric constraint (MFCC) for AD analysis. First, the similarity is calculated for each modality (i.e. VBM-MRI or FDG-PET) individually by random forest strategy, which can extract pairwise similarity measures for multiple modalities. Then the group sparsity regularization term and the sample similarity constraint regularization term are used to constrain the objective function to conduct feature selection from multiple modalities. Finally, the multi-kernel support vector machine (MK-SVM) is used to fuse the features selected from different models for final classification. The experimental results on the Alzheimer's Disease Neuroimaging Initiative (ADNI) show that the proposed method has better classification performance than the start-of-the-art multimodality-based methods. Specifically, we achieved higher accuracy and area under the curve (AUC) for AD versus normal controls (NC), MCI versus NC, and MCI converters (MCI-C) versus MCI non-converters (MCI-NC) on ADNI datasets. Therefore, the proposed model not only outperforms the traditional method in terms of AD/MCI classification, but also discovers the characteristics associated with the disease, demonstrating its promise for improving disease-related mechanistic understanding.

Leveraging Coupled Interaction for Multimodal Alzheimer's Disease Diagnosis.

As the population becomes older worldwide, accurate computer-aided diagnosis for Alzheimer's disease (AD) in the early stage has been regarded as a crucial step for neurodegeneration care in recent years. Since it extracts the low-level features from the neuroimaging data, previous methods regarded this computer-aided diagnosis as a classification problem that ignored latent featurewise relation. However, it is known that multiple brain regions in the human brain are anatomically and functionally interlinked according to the current neuroscience perspective. Thus, it is reasonable to assume that the extracted features from different brain regions are related to each other to some extent. Also, the complementary information between different neuroimaging modalities could benefit multimodal fusion. To this end, we consider leveraging the coupled interactions in the feature level and modality level for diagnosis in this paper. First, we propose capturing the feature-level coupled interaction using a coupled feature representation. Then, to model the modality-level coupled interaction, we present two novel methods: 1) the coupled boosting (CB) that models the correlation of pairwise coupled-diversity on both inconsistently and incorrectly classified samples between different modalities and 2) the coupled metric ensemble (CME) that learns an informative feature projection from different modalities by integrating the intrarelation and interrelation of training samples. We systematically evaluated our methods with the AD neuroimaging initiative data set. By comparison with the baseline learning-based methods and the state-of-the-art methods that are specially developed for AD/MCI (mild cognitive impairment) diagnosis, our methods achieved the best performance with accuracy of 95.0% and 80.7% (CB), 94.9% and 79.9% (CME) for AD/NC (normal control), and MCI/NC identification, respectively.

Classifying cognitive impairment based on the spatial heterogeneity of cerebral blood flow images.

The spatial coefficient of variation (sCoV) of arterial spin-labeled (ASL) MRI can index cerebral blood flow spatial heterogeneity. This metric reflects delayed blood delivery-seen as a hyperintense ASL signal juxtaposed by hypointense regions. To investigate the use of ASL-sCoV in the classification of cognitively unimpaired (CU), mild cognitive impairment (MCI), and Alzheimer's disease (AD) cohorts. Prospective/cohort. Baseline ASL images from AD neuroimaging initiative dataset in three groups of CU, MCI, and AD (N = 258). Pulsed ASL (PICORE QT2) images were acquired on 3 T Siemens systems (TE/TR = 12/3400 msec, TI1/2 = 700/1900 msec). ASL-sCoV was calculated in temporal, parietal, occipital, and frontal lobes as well as whole gray matter. The primary analysis used an analysis of covariance to investigate sCoV and cognitive group (CU, MCI, AD) associations. We also evaluated the repeatability of sCoV by calculating within-subject agreement in a subgroup of CU participants with a repeat ASL. The secondary analyses assessed ventricular volume, amyloid burden, glucose uptake, ASL-sCoV, and regional CBF as cognitive group classifiers using logistic regression models and receiver operating characteristic analyses. We found that global and temporal lobe sCoV differed between cognitive groups (P = 0.006). Post-hoc tests showed that temporal lobe sCoV was lower in CU than in MCI (Cohen's d = -0.36) or AD (Cohen's d = -1.36). We found that sCoV was moderately repeatable in CU (intersession intraclass correlation = 0.50; intrasession intraclass correlation = 0.88). Subsequent logistic regression analyses revealed that temporal lobe sCoV and amyloid uptake classified CU vs. MCI (P < 0.01; accuracy = 78%). Temporal lobe sCoV, amyloid, and glucose uptake classified CU vs. AD (P < 0.01; accuracy = 97%); glucose uptake significantly classified MCI vs. AD (P < 0.01; accuracy = 85%). We showed that ASL spatial heterogeneity can be used alongside AD neuroimaging markers to distinguish cognitive groups, in particular, cognitively unimpaired from cognitively impaired individuals. 2 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2019;50:858-867.

Independent and Reproducible Hippocampal Radiomic Biomarkers for Multisite Alzheimer's Disease: Diagnosis, Longitudinal Progress and Biological Basis

Hippocampal morphological change is one of the main hallmarks of Alzheimer's disease (AD). However, whether hippocampal radiomic features are reproducible and robust as predictors of progression from mild cognitive impairment (MCI) to AD dementia and provide a neurobiological foundation remains unclear. The primary aim of this study was to verify whether hippocampal radiomic features are robust MRI markers for AD. Multivariate classifier-based SVM analysis provided individual-level predictions for distinguishing AD patients (N=261) from normal controls (NCs; N=231) with accuracy=88.21% with inter-site cross-validation. Further analyses of a large, independent ADNI dataset (N=1228) reinforced these findings. In mild cognitive impairment (MCI) groups, a systemic analysis demonstrated that the identified features were significantly associated with clinical features (e.g., apolipoprotein E (APOE) genotype, polygenic risk scores, cerebrospinal fluid (CSF) Aβ, CSF Tau), and longitudinal changes in cognition ability; more importantly, the radiomic features have a consistently altered pattern with changes in the MMSE scores over 5 years of follow-up. These comprehensive results suggest that hippocampal radiomic features can serve as a robust biomarker for clinical applications in AD/MCI, further provide evidence for predicting whether MCI subject would convert to AD or not based on hippocampus, without doubt has significantly meaning for early diagnosis of AD/MCI. Funding Statement: This work was partially supported by the National Key Research and Development Program of China (grant no. 2016YFC1305904), the Strategic Priority Research Program (B) of the Chinese Academy of Sciences (grant no. XDB32020200), the National Natural Science Foundation of China (grant nos. 81871438, 81901101, 61633018, 81571062, 81400890, 61431012), and the Primary Research & Development Plan of Shandong Province (grant no. 2017GGX10112). Data collection and sharing for this project was funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH 12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and generous contributions from AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research provide funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. The ADNI data are disseminated by the Laboratory for Neuro-Imaging at the University of Southern California. Declaration of Interests: This study includes the databases which were approved by the Medical Ethics Committee of local Hospitals.

Correlation-Aware Sparse and Low-Rank Constrained Multi-Task Learning for Longitudinal Analysis of Alzheimer's Disease.

Alzheimer's disease (AD), as a severe neurodegenerative disease, is now attracting more and more researchers' attention in the healthcare. With the development of magnetic resonance imaging (MRI), the neuroimaging-based longitudinal analysis is gradually becoming an important research direction to understand and trace the process of the AD. In addition, regression analysis has been commonly adopted in the AD pattern analysis and progression prediction. However, most existing methods assume that all input features are equally related to the output variables, which ignore the difference in terms of the correlation. In this paper, we proposed a novel multi-task learning formulation, which considers a correlation-aware sparse and low-rank constrained regularization, for accurately predicting the cognitive scores of the patients at different time points and identifying the most predictive biomarkers. In addition, an efficient iterative algorithm is developed to optimize the proposed non-smooth convex objective formulation. We also have performed experiments using data from the AD neuroimaging initiative dataset to evaluate the proposed optimization formulation. Especially, we will predict cognitive scores of multiple time points through the baseline MRI features. The results not only indicate the rationality and correctness of the proposed method for predicting disease progression but also identify some stable and important MRI features that are consistent with the previous research.

A Deep Learning Model to Predict a Diagnosis of Alzheimer Disease by Using 18F-FDG PET of the Brain.

Purpose To develop and validate a deep learning algorithm that predicts the final diagnosis of Alzheimer disease (AD), mild cognitive impairment, or neither at fluorine 18 (18F) fluorodeoxyglucose (FDG) PET of the brain and compare its performance to that of radiologic readers. Materials and Methods Prospective 18F-FDG PET brain images from the Alzheimer's Disease Neuroimaging Initiative (ADNI) (2109 imaging studies from 2005 to 2017, 1002 patients) and retrospective independent test set (40 imaging studies from 2006 to 2016, 40 patients) were collected. Final clinical diagnosis at follow-up was recorded. Convolutional neural network of InceptionV3 architecture was trained on 90% of ADNI data set and tested on the remaining 10%, as well as the independent test set, with performance compared to radiologic readers. Model was analyzed with sensitivity, specificity, receiver operating characteristic (ROC), saliency map, and t-distributed stochastic neighbor embedding. Results The algorithm achieved area under the ROC curve of 0.98 (95% confidence interval: 0.94, 1.00) when evaluated on predicting the final clinical diagnosis of AD in the independent test set (82% specificity at 100% sensitivity), an average of 75.8 months prior to the final diagnosis, which in ROC space outperformed reader performance (57% [four of seven] sensitivity, 91% [30 of 33] specificity; P < .05). Saliency map demonstrated attention to known areas of interest but with focus on the entire brain. Conclusion By using fluorine 18 fluorodeoxyglucose PET of the brain, a deep learning algorithm developed for early prediction of Alzheimer disease achieved 82% specificity at 100% sensitivity, an average of 75.8 months prior to the final diagnosis. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Larvie in this issue.

Reconstructing subject-specific effect maps

Predictive models allow subject-specific inference when analyzing disease related alterations in neuroimaging data. Given a subject's data, inference can be made at two levels: global, i.e. identifiying condition presence for the subject, and local, i.e. detecting condition effect on each individual measurement extracted from the subject's data. While global inference is widely used, local inference, which can be used to form subject-specific effect maps, is rarely used because existing models often yield noisy detections composed of dispersed isolated islands. In this article, we propose a reconstruction method, named RSM, to improve subject-specific detections of predictive modeling approaches and in particular, binary classifiers. RSM specifically aims to reduce noise due to sampling error associated with using a finite sample of examples to train classifiers. The proposed method is a wrapper-type algorithm that can be used with different binary classifiers in a diagnostic manner, i.e. without information on condition presence. Reconstruction is posed as a Maximum-A-Posteriori problem with a prior model whose parameters are estimated from training data in a classifier-specific fashion. Experimental evaluation is performed on synthetically generated data and data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Results on synthetic data demonstrate that using RSM yields higher detection accuracy compared to using models directly or with bootstrap averaging. Analyses on the ADNI dataset show that RSM can also improve correlation between subject-specific detections in cortical thickness data and non-imaging markers of Alzheimer's Disease (AD), such as the Mini Mental State Examination Score and Cerebrospinal Fluid amyloid-β levels. Further reliability studies on the longitudinal ADNI dataset show improvement on detection reliability when RSM is used.

Brain-Wide Genome-Wide Association Study for Alzheimer's Disease via Joint Projection Learning and Sparse Regression Model.

Brain-wide and genome-wide association (BW-GWA) study is presented in this paper to identify the associations between the brain imaging phenotypes (i.e., regional volumetric measures) and the genetic variants [i.e., single nucleotide polymorphism (SNP)] in Alzheimer's disease (AD). The main challenges of this study include the data heterogeneity, complex phenotype-genotype associations, high-dimensional data (e.g., thousands of SNPs), and the existence of phenotype outliers. Previous BW-GWA studies, while addressing some of these challenges, did not consider the diagnostic label information in their formulations, thus limiting their clinical applicability. To address these issues, we present a novel joint projection and sparse regression model to discover the associations between the phenotypes and genotypes. Specifically, to alleviate the negative influence of data heterogeneity, we first map the genotypes into an intermediate imaging-phenotype-like space. Then, to better reveal the complex phenotype-genotype associations, we project both the mapped genotypes and the original imaging phenotypes into a diagnostic-label-guided joint feature space, where the intraclass projected points are constrained to be close to each other. In addition, we use l2,1-norm minimization on both the regression loss function and the transformation coefficient matrices, to reduce the effect of phenotype outliers and also to encourage sparse feature selections of both the genotypes and phenotypes. We evaluate our method using AD neuroimaging initiative dataset, and the results show that our proposed method outperforms several state-of-the-art methods in term of the average root-mean-square error of genome-to-phenotype predictions. Besides, the associated SNPs and brain regions identified in this study have also been shown in the previous AD-related studies, thus verifying the effectiveness and potential of our proposed method in AD pathogenesis study.