Faculty Bibliography
Searched for:
in-biosketch:true
person:sadowm01
Total Results:
162
Performing Sparse Regularization and Dimension Reduction Simultaneously in Multimodal Data Fusion
Collecting multiple modalities of neuroimaging data on the same subject is increasingly becoming the norm in clinical practice and research. Fusing multiple modalities to find related patterns is a challenge in neuroimaging analysis. Canonical correlation analysis (CCA) is commonly used as a symmetric data fusion technique to find related patterns among multiple modalities. In CCA-based data fusion, principal component analysis (PCA) is frequently applied as a preprocessing step to reduce data dimension followed by CCA on dimension-reduced data. PCA, however, does not differentiate between informative voxels from non-informative voxels in the dimension reduction step. Sparse PCA (sPCA) extends traditional PCA by adding sparse regularization that assigns zero weights to non-informative voxels. In this study, sPCA is incorporated into CCA-based fusion analysis and applied on neuroimaging data. A cross-validation method is developed and validated to optimize the parameters in sPCA. Different simulations are carried out to evaluate the improvement by introducing sparsity constraint to PCA. Four fusion methods including sPCA+CCA, PCA+CCA, parallel ICA and sparse CCA were applied on structural and functional magnetic resonance imaging data of mild cognitive impairment subjects and normal controls. Our results indicate that sPCA significantly can reduce the impact of non-informative voxels and lead to improved statistical power in uncovering disease-related patterns by a fusion analysis.
PMCID:6618346
PMID: 31333396
ISSN: 1662-4548
CID: 5864642
Robust Motion Regression of Resting-State Data Using a Convolutional Neural Network Model
Resting-state functional magnetic resonance imaging (rs-fMRI) based on the blood-oxygen-level-dependent (BOLD) signal has been widely used in healthy individuals and patients to investigate brain functions when the subjects are in a resting or task-negative state. Head motion considerably confounds the interpretation of rs-fMRI data. Nuisance regression is commonly used to reduce motion-related artifacts with six motion parameters estimated from rigid-body realignment as regressors. To further compensate for the effect of head movement, the first-order temporal derivatives of motion parameters and squared motion parameters were proposed previously as possible motion regressors. However, these additional regressors may not be sufficient to model the impact of head motion because of the complexity of motion artifacts. In addition, while using more motion-related regressors could explain more variance in the data, the neural signal may also be removed with increasing number of motion regressors. To better model how in-scanner motion affects rs-fMRI data, a robust and automated convolutional neural network (CNN) model is developed in this study to obtain optimal motion regressors. The CNN network consists of two temporal convolutional layers and the output from the network are the derived motion regressors used in the following nuisance regression. The temporal convolutional layer in the network can non-parametrically model the prolonged effect of head motion. The set of regressors derived from the neural network is compared with the same number of regressors used in a traditional nuisance regression approach. It is demonstrated that the CNN-derived regressors can more effectively reduce motion-related artifacts.
PMCID:6482337
PMID: 31057348
ISSN: 1662-4548
CID: 5134372
Quantitative 18F-AV1451 Brain Tau PET Imaging in Cognitively Normal Older Adults, Mild Cognitive Impairment, and Alzheimer's Disease Patients
Recent developments of tau Positron Emission Tomography (PET) allows assessment of regional neurofibrillary tangles (NFTs) deposition in human brain. Among the tau PET molecular probes, 18F-AV1451 is characterized by high selectivity for pathologic tau aggregates over amyloid plaques, limited non-specific binding in white and gray matter, and confined off-target binding. The objectives of the study are (1) to quantitatively characterize regional brain tau deposition measured by 18F-AV1451 PET in cognitively normal older adults (CN), mild cognitive impairment (MCI), and AD participants; (2) to evaluate the correlations between cerebrospinal fluid (CSF) biomarkers or Mini-Mental State Examination (MMSE) and 18F-AV1451 PET standardized uptake value ratio (SUVR); and (3) to evaluate the partial volume effects on 18F-AV1451 brain uptake. Methods: The study included total 115 participants (CN = 49, MCI = 58, and AD = 8) from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Preprocessed 18F-AV1451 PET images, structural MRIs, and demographic and clinical assessments were downloaded from the ADNI database. A reblurred Van Cittertiteration method was used for voxelwise partial volume correction (PVC) on PET images. Structural MRIs were used for PET spatial normalization and region of interest (ROI) definition in standard space. The parametric images of 18F-AV1451 SUVR relative to cerebellum were calculated. The ROI SUVR measurements from PVC and non-PVC SUVR images were compared. The correlation between ROI 18F-AV1451 SUVR and the measurements of MMSE, CSF total tau (t-tau), and phosphorylated tau (p-tau) were also assessed. Results: 18F-AV1451 prominently specific binding was found in the amygdala, entorhinal cortex, parahippocampus, fusiform, posterior cingulate, temporal, parietal, and frontal brain regions. Most regional SUVRs showed significantly higher uptake of 18F-AV1451 in AD than MCI and CN participants. SUVRs of small regions like amygdala, entorhinal cortex and parahippocampus were statistically improved by PVC in all groups (p < 0.01). Although there was an increasing tendency of 18F-AV-1451 SUVRs in MCI group compared with CN group, no significant difference of 18F-AV1451 deposition was found between CN and MCI brains with or without PVC (p > 0.05). Declined MMSE score was observed with increasing 18F-AV1451 binding in amygdala, entorhinal cortex, parahippocampus, and fusiform. CSF p-tau was positively correlated with 18F-AV1451 deposition. PVC improved the results of 18F-AV-1451 tau deposition and correlation studies in small brain regions. Conclusion: The typical deposition of 18F-AV1451 tau PET imaging in AD brain was found in amygdala, entorhinal cortex, fusiform and parahippocampus, and these regions were strongly associated with cognitive impairment and CSF biomarkers. Although more deposition was observed in MCI group, the 18F-AV-1451 PET imaging could not differentiate the MCI patients from CN population. More tau deposition related to decreased MMSE score and increased level of CSF p-tau, especially in ROIs of amygdala, entorhinal cortex and parahippocampus. PVC did improve the results of tau deposition and correlation studies in small brain regions and suggest to be routinely used in 18F-AV1451 tau PET quantification.
PMCID:6530456
PMID: 31156534
ISSN: 1664-2295
CID: 5864632
18F-florbetapir Positron Emission Tomography-determined Cerebral beta-Amyloid Deposition and Neurocognitive Performance after Cardiac Surgery
BACKGROUND:Amyloid deposition is a potential contributor to postoperative cognitive dysfunction. The authors hypothesized that 6-week global cortical amyloid burden, determined by F-florbetapir positron emission tomography, would be greater in those patients manifesting cognitive dysfunction at 6 weeks postoperatively. METHODS:Amyloid deposition was evaluated in cardiac surgical patients at 6 weeks (n = 40) and 1 yr (n = 12); neurocognitive function was assessed at baseline (n = 40), 6 weeks (n = 37), 1 yr (n = 13), and 3 yr (n = 9). The association of 6-week amyloid deposition with cognitive dysfunction was assessed by multivariable regression, accounting for age, years of education, and baseline cognition. Differences between the surgical cohort with cognitive deficit and the Alzheimer's Disease Neuroimaging Initiative cohorts (normal and early/late mild cognitive impairment) was assessed, adjusting for age, education, and apolipoprotein E4 genotype. RESULTS:The authors found that 6-week abnormal global cortical amyloid deposition was not associated with cognitive dysfunction (13 of 37, 35%) at 6 weeks postoperatively (median standard uptake value ratio [interquartile range]: cognitive dysfunction 0.92 [0.89 to 1.07] vs. 0.98 [0.93 to 1.05]; P = 0.455). In post hoc analyses, global cortical amyloid was also not associated with cognitive dysfunction at 1 or 3 yr postoperatively. Amyloid deposition at 6 weeks in the surgical cohort was not different from that in normal Alzheimer's Disease Neuroimaging Initiative subjects, but increased over 1 yr in many areas at a rate greater than in controls. CONCLUSIONS:In this study, postoperative cognitive dysfunction was not associated with 6-week cortical amyloid deposition. The relationship between cognitive dysfunction and regional amyloid burden and the rate of postoperative amyloid deposition merit further investigation.
PMCID:5849499
PMID: 29389750
ISSN: 1528-1175
CID: 2994312
Biomarker pattern of ARIA-E participants in phase 3 randomized clinical trials with bapineuzumab
OBJECTIVE:To evaluate whether amyloid-related imaging abnormalities with edema/effusion (ARIA-E) observed in bapineuzumab clinical trials was associated with specific biomarker patterns. METHODS:Bapineuzumab, an anti-β-amyloid monoclonal antibody, was evaluated in patients with mild to moderate Alzheimer disease. Amyloid PET imaging, CSF biomarkers, or volumetric MRI (vMRI) were assessed. RESULTS:. CONCLUSIONS:Baseline biomarkers largely do not predict risk for developing ARIA-E. ARIA-E was associated with significant longitudinal changes in several biomarkers, with larger reductions in amyloid PET and CSF p-tau and t-tau concentrations, and paradoxically greater hippocampal volume reduction and ventricular enlargement, suggesting that ARIA-E in bapineuzumab-treated cases may be related to increased Aβ efflux from the brain and affecting downstream pathogenic processes.
PMID: 29429971
ISSN: 1526-632x
CID: 3256922
Proceedings of the National Academy of Sciences of the United States of America (PNAS). 2018:115(8):1697-1706.DOI: 10.1073/pnas.1715554115
Identification of genetic risk factors in the Chinese population implicates a role of immune system in Alzheimer's disease pathogenesis
Alzheimer's disease (AD) is a leading cause of mortality among the elderly. We performed a whole-genome sequencing study of AD in the Chinese population. In addition to the variants identified in or around the APOE locus (sentinel variant rs73052335, P = 1.44 × 10-14), two common variants, GCH1 (rs72713460, P = 4.36 × 10-5) and KCNJ15 (rs928771, P = 3.60 × 10-6), were identified and further verified for their possible risk effects for AD in three small non-Asian AD cohorts. Genotype-phenotype analysis showed that KCNJ15 variant rs928771 affects the onset age of AD, with earlier disease onset in minor allele carriers. In addition, altered expression level of the KCNJ15 transcript can be observed in the blood of AD subjects. Moreover, the risk variants of GCH1 and KCNJ15 are associated with changes in their transcript levels in specific tissues, as well as changes of plasma biomarkers levels in AD subjects. Importantly, network analysis of hippocampus and blood transcriptome datasets suggests that the risk variants in the APOE, GCH1, and KCNJ15 loci might exert their functions through their regulatory effects on immune-related pathways. Taking these data together, we identified common variants of GCH1 and KCNJ15 in the Chinese population that contribute to AD risk. These variants may exert their functional effects through the immune system.
PMID: 29432188
ISSN: 1091-6490
CID: 5864612
Proceedings of the National Academy of Sciences of the United States of America (PNAS). 2018:115(7):1481-1486.DOI: 10.1073/pnas.1719747115
Statistical tests and identifiability conditions for pooling and analyzing multisite datasets
When sample sizes are small, the ability to identify weak (but scientifically interesting) associations between a set of predictors and a response may be enhanced by pooling existing datasets. However, variations in acquisition methods and the distribution of participants or observations between datasets, especially due to the distributional shifts in some predictors, may obfuscate real effects when datasets are combined. We present a rigorous statistical treatment of this problem and identify conditions where we can correct the distributional shift. We also provide an algorithm for the situation where the correction is identifiable. We analyze various properties of the framework for testing model fit, constructing confidence intervals, and evaluating consistency characteristics. Our technical development is motivated by Alzheimer's disease (AD) studies, and we present empirical results showing that our framework enables harmonizing of protein biomarkers, even when the assays across sites differ. Our contribution may, in part, mitigate a bottleneck that researchers face in clinical research when pooling smaller sized datasets and may offer benefits when the subjects of interest are difficult to recruit or when resources prohibit large single-site studies.
PMCID:5816202
PMID: 29386387
ISSN: 1091-6490
CID: 3257392
Dual-Model Radiomic Biomarkers Predict Development of Mild Cognitive Impairment Progression to Alzheimer's Disease
Predicting progression of mild cognitive impairment (MCI) to Alzheimer's disease (AD) is clinically important. In this study, we propose a dual-model radiomic analysis with multivariate Cox proportional hazards regression models to investigate promising risk factors associated with MCI conversion to AD. T1 structural magnetic resonance imaging (MRI) and 18F-Fluorodeoxyglucose (FDG) positron emission tomography (PET) data, from the AD Neuroimaging Initiative database, were collected from 131 patients with MCI who converted to AD within 3 years and 132 patients with MCI without conversion within 3 years. These subjects were randomly partition into 70% training dataset and 30% test dataset with multiple times. We fused MRI and PET images by wavelet method. In a subset of subjects, a group comparison was performed using a two-sample t-test to determine regions of interest (ROIs) associated with MCI conversion. 172 radiomic features from ROIs for each individual were established using a published radiomics tool. Finally, L1-penalized Cox model was constructed and Harrell's C index (C-index) was used to evaluate prediction accuracy of the model. To evaluate the efficacy of our proposed method, we used a same analysis framework to evaluate MRI and PET data separately. We constructed prognostic Cox models with: clinical data, MRI images, PET images, fused MRI/PET images, and clinical variables and fused MRI/PET images in combination. The experimental results showed that captured ROIs significantly associated with conversion to AD, such as gray matter atrophy in the bilateral hippocampus and hypometabolism in the temporoparietal cortex. Imaging model (MRI/PET/fused) provided significant enhancement in prediction of conversion compared to clinical models, especially the fused-modality Cox model. Moreover, the combination of fused-modality imaging and clinical variables resulted in the greatest accuracy of prediction. The average C-index for the clinical/MRI/PET/fused/combined model in the test dataset was 0.69, 0.73, 0.73 and 0.75, and 0.78, respectively. These results suggested that a combination of radiomic analysis and Cox model analyses could be used successfully in survival analysis and may be powerful tools for personalized precision medicine patients with potential to undergo conversion from MCI to AD.
PMCID:6338093
PMID: 30686995
ISSN: 1662-4548
CID: 5134342
Pattern Discovery in Brain Imaging Genetics via SCCA Modeling with a Generic Non-convex Penalty
Brain imaging genetics intends to uncover associations between genetic markers and neuroimaging quantitative traits. Sparse canonical correlation analysis (SCCA) can discover bi-multivariate associations and select relevant features, and is becoming popular in imaging genetic studies. The L1-norm function is not only convex, but also singular at the origin, which is a necessary condition for sparsity. Thus most SCCA methods impose [Formula: see text]-norm onto the individual feature or the structure level of features to pursuit corresponding sparsity. However, the [Formula: see text]-norm penalty over-penalizes large coefficients and may incurs estimation bias. A number of non-convex penalties are proposed to reduce the estimation bias in regression tasks. But using them in SCCA remains largely unexplored. In this paper, we design a unified non-convex SCCA model, based on seven non-convex functions, for unbiased estimation and stable feature selection simultaneously. We also propose an efficient optimization algorithm. The proposed method obtains both higher correlation coefficients and better canonical loading patterns. Specifically, these SCCA methods with non-convex penalties discover a strong association between the APOE e4 rs429358 SNP and the hippocampus region of the brain. They both are Alzheimer's disease related biomarkers, indicating the potential and power of the non-convex methods in brain imaging genetics.
PMCID:5656688
PMID: 29070790
ISSN: 2045-2322
CID: 3257412
Translational Control of APP Expression for Alzheimer Disease Therapy [Editorial]
ORIGINAL:0012875
ISSN: 2573-6051
CID: 3257472
