Shannon L. Risacher

Sparse multi-task regression and feature selection to identify brain imaging predictors for memory performance

Alzheimers disease (AD) is a neurodegenerative disorder characterized by progressive impairment of memory and other cognitive functions, which makes regression analysis a suitable model to study whether neuroimaging measures can help predict memory performance and track the progression of AD. Existing memory performance prediction methods via regression, however, do not take into account either the interconnected structures within imaging data or those among memory scores, which inevitably restricts their predictive capabilities. To bridge this gap, we propose a novel Sparse Multi-tAsk Regression and feaTure selection (SMART) method to jointly analyze all the imaging and clinical data under a single regression framework and with shared underlying sparse representations. Two convex regularizations are combined and used in the model to enable sparsity as well as facilitate multi-task learning. The effectiveness of the proposed method is demonstrated by both clearly improved prediction performances in all empirical test cases and a compact set of selected RAVLT-relevant MRI predictors that accord with prior studies.

Cortical surface biomarkers for predicting cognitive outcomes using group ℓ2,1 norm

Regression models have been widely studied to investigate the prediction power of neuroimaging measures as biomarkers for inferring cognitive outcomes in the Alzheimers disease study. Most of these models ignore the interrelated structures either within neuroimaging measures or between cognitive outcomes, and thus may have limited power to yield optimal solutions. To address this issue, we propose to use a new sparse multitask learning model called Group-Sparse Multi-task Regression and Feature Selection (G-SMuRFS) and demonstrate its effectiveness by examining the predictive power of detailed cortical thickness measures toward 3 types of cognitive scores in a large cohort. G-SMuRFS proposes a group-level l2,1-norm strategy to group relevant features together in an anatomically meaningful manner and use this prior knowledge to guide the learning process. This approach also takes into account the correlation among cognitive outcomes for building a more appropriate predictive model. Compared with traditional methods, G-SMuRFS not only demonstrates a superior performance but also identifies a small set of surface markers that are biologically meaningful.

Identification of associations between genotypes and longitudinal phenotypes via temporally-constrained group sparse canonical correlation analysis

Motivation: Neuroimaging genetics identifies the relationships between genetic variants (i.e., the single nucleotide polymorphisms) and brain imaging data to reveal the associations from genotypes to phenotypes. So far, most existing machine‐learning approaches are widely used to detect the effective associations between genetic variants and brain imaging data at one time‐point. However, those associations are based on static phenotypes and ignore the temporal dynamics of the phenotypical changes. The phenotypes across multiple time‐points may exhibit temporal patterns that can be used to facilitate the understanding of the degenerative process. In this article, we propose a novel temporally constrained group sparse canonical correlation analysis (TGSCCA) framework to identify genetic associations with longitudinal phenotypic markers. Results: The proposed TGSCCA method is able to capture the temporal changes in brain from longitudinal phenotypes by incorporating the fused penalty, which requires that the differences between two consecutive canonical weight vectors from adjacent time‐points should be small. A new efficient optimization algorithm is designed to solve the objective function. Furthermore, we demonstrate the effectiveness of our algorithm on both synthetic and real data (i.e., the Alzheimers Disease Neuroimaging Initiative cohort, including progressive mild cognitive impairment, stable MCI and Normal Control participants). In comparison with conventional SCCA, our proposed method can achieve strong associations and discover phenotypic biomarkers across multiple time‐points to guide disease‐progressive interpretation. Availability and implementation: The Matlab code is available at https://sourceforge.net/projects/ibrain‐cn/files/. Contact: dqzhang@nuaa.edu.cn or shenli@iu.edu

Associations of the Top 20 Alzheimer Disease Risk Variants With Brain Amyloidosis

Importance Late-onset Alzheimer disease (AD) is highly heritable. Genome-wide association studies have identified more than 20 AD risk genes. The precise mechanism through which many of these genes are associated with AD remains unknown. Objective To investigate the association of the top 20 AD risk variants with brain amyloidosis. Design, Setting, and Participants This study analyzed the genetic and florbetapir F 18 data from 322 cognitively normal control individuals, 496 individuals with mild cognitive impairment, and 159 individuals with AD dementia who had genome-wide association studies and 18F-florbetapir positron emission tomographic data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI), a prospective, observational, multisite tertiary center clinical and biomarker study. This ongoing study began in 2005. Main Outcomes and Measures The study tested the association of AD risk allele carrier status (exposure) with florbetapir mean standard uptake value ratio (outcome) using stepwise multivariable linear regression while controlling for age, sex, and apolipoprotein E &egr;4 genotype. The study also reports on an exploratory 3-dimensional stepwise regression model using an unbiased voxelwise approach in Statistical Parametric Mapping 8 with cluster and significance thresholds at 50 voxels and uncorrected Pu2009<u2009.01. Results This study included 977 participants (mean [SD] age, 74 [7.5] years; 535 [54.8%] male and 442 [45.2%] female) from the ADNI-1, ADNI-2, and ADNI–Grand Opportunity. The adenosine triphosphate–binding cassette subfamily A member 7 (ABCA7) gene had the strongest association with amyloid deposition (&khgr;2u2009=u20098.38, false discovery rate–corrected P < .001), after apolioprotein E &egr;4. Significant associations were found between ABCA7 in the asymptomatic and early symptomatic disease stages, suggesting an association with rapid amyloid accumulation. The fermitin family homolog 2 (FERMT2) gene had a stage-dependent association with brain amyloidosis (FERMT2 × diagnosis &khgr;2u2009=u20093.53, false discovery rate–corrected P =u2009.05), which was most pronounced in the mild cognitive impairment stage. Conclusions and Relevance This study found an association of several AD risk variants with brain amyloidosis. The data also suggest that AD genes might differentially regulate AD pathologic findings across the disease stages.

Network-based genome wide study of hippocampal imaging phenotype in Alzheimer's Disease to identify functional interaction modules

Identification of functional modules from biological network is a promising approach to enhance the statistical power of genome-wide association study (GWAS) and improve biological interpretation for complex diseases. The precise functions of genes are highly relevant to tissue context, while a majority of module identification studies are based on tissuefree biological networks that lacks phenotypic specificity. In this study, we propose a module identification method that maps the GWAS results of an imaging phenotype onto the corresponding tissue-specific functional interaction network by applying a machine learning framework. Ridge regression and support vector machine (SVM) models are constructed to re-prioritize GWAS results, followed by exploring hippocampus-relevant modules based on top predictions using GWAS top findings. We also propose a GWAS top-neighbor-based module identification approach and compare it with Ridge and SVM based approaches. Modules conserving both tissue specificity and GWAS discoveries are identified, showing the promise of the proposal method for providing insight into the mechanism of complex diseases.

Pairwise gene-protein association analysis on cerebrospinal fluid proteomics in the ADNI-1 cohort

numerous genetic studies to be undertaken in Alzheimer’s disease (AD) populations resulting in the identification of new risk factors. The importance of both genetics and environment in brain function is well known, as is the role of neuroimaging in revealing brain dysfunction, neuronal loss and neocortical Ab burden (NAB) through positron emission tomography (PET) using a radioactive tracer (e.g. PiB). Thus, the synergy of integrating genetics with brain imaging carries clear advantages. These quantitative trait (QT) analyses can be further extended to include phenotypes such as clinical and cognitive standardized assessments and other disease biomarkers. Here, we report on gene associations, both independent and combinatorial, with AD risk and QTs in the Australian Imaging, Biomarkers and Lifestyle (AIBL) study cohort. Methods: This study reports on the recently completed screen of candidate genes in the Australian Imaging, Biomarkers and Lifestyle (AIBL) study (942 individuals). Single Nucleotide Polymorphisms (SNPs) were selected from the top candidate genes identified in AD association and QT studies (specific hits and fine-mapping). Genotyping was via a custom Illumina GoldenGate assay, with failed SNPs and additional supplemental SNPs genotyped using the OpenArray Platform. Genotype data that passed quality control were then analysed with respect to clinical classification of disease and QT traits including, but not limited to; plasma Ab levels, hippocampal volume, NAB, and measures of cognitive performance. Novel machine learning algorithms were utilised to identify combinations of SNPs that classified AD or allowed for the prediction of high NAB (defined as a greater than 1.3 Standardized Uptake Value Ratio (SUVR) determined via PiB-PET). Results: A total of over 1525 SNPs passed quality control and 935 individuals (call rate greater than 98%) were included in analyses. Single marker and haplotype associations with AD risk and QTs were identified using GoldenHelix SVS7 software. Novel ML models have identified a combination of SNPs able to classify AD or high NAB. Conclusions: The use of a quantitative phenotype in combination with genetic studies provides many advantages over a case-control design, both in terms of power and in terms of physiological understanding of the underlying cognitive and pathological processes.