Department of Biostatistics
Neurostatistics Working Group

2013 - 2014

Coordinators: Dr. Rebecca Betensky and Dr. Josephine Asafu-Adjei

Schedule: Wednesdays, 12:30-1:30 p.m.
HSPH2, Room 426 (unless otherwise notified)

Contract All | Expand All
Seminar Description
This working group provides a forum for presentation and discussion of completed, ongoing, or planned statistical analyses of neurological data. Such data include, for example, in vivo human brain images (anatomic, functional and spectroscopic magnetic resonance imaging), gene expression studies of human and non-human animal brain tissue (brightfield and immunofluorescence microscopy, DNA microarrays, laser micro-dissection), in vivo micro-dialysis, clinical trials data for a variety of neurologic diseases, and genetic data from family studies. Non-statistical presentations of neurological, psychiatric and technological background material will also be included. Through this seminar, statisticians will gain exposure to the statistical issues that arise in the broad field of neurology and brain imaging psychiatry and to the diverse ongoing research in this area throughout Harvard and the world. A main goal of the seminar is to stimulate statistical interest in neuroscience and neurology and to develop strategies for collaboration within these fields.


July 30

Folefac Atem, Ph.D.
Research Fellow, Department of Biostatistics, Harvard School of Public Health

"Estimation of The Linear Model with Right Censored Covariates"
ABSTRACT: See attached
and

Josephine K. Asafu-Adjei, Ph.D.
Research Fellow, Department of Biostatistics, Harvard School of Public Health

"Bayesian Variable Selection Methods for Matched Case-Control Studies"
ABSTRACT: The use of matched designs in case-control studies can result in substantial improvements in efficiency and statistical power. However, it is quite common for studies, particularly those dealing with high dimensional data, to ignore matching when applying variable selection techniques, which can reduce precision and validity. We will present our work in developing a new approach to account for matching in the context of Bayesian variable selection methods, which effectively handle high dimensional data settings.

September 25

Kyu Ha Lee, M.D., Ph.D.
Research Fellow, Department of Biostatistics, Harvard School of Public Health

"Bayesian Semi-parametric Analysis of Semi-competing Risks Data: Investigating Hospital Readmission after a Pancreatic Cancer Diagnosis"
ABSTRACT: In the U.S. the Centers for Medicare and Medicaid Services uses 30-day readmission, following hospitalization, as a proxy outcome to monitor quality of care. These efforts generally focus on treatable health conditions that have good prognoses, such as pneumonia and heart failure. Expanding quality of care systems to monitor conditions with poor prognosis, such as pancreatic cancer, is challenging because of the non-trivial force of mortality; 30-day mortality for pancreatic cancer is approximately 30%. In the statistical literature, the study of a non-terminal event, such as readmission, whose observation is subject to a terminal event, such as death, is known as the 'semi-competing risks' problem. Existing statistical methods focus almost exclusively on estimation of regression parameters. Methods are sparse or non-existent, however, when scientific interest also lies in the characterization of the dependence structure between the terminal and non-terminal events or in the prediction of either event for a patient with a given covariate profile. In this paper we propose a Bayesian semi-parametric regression framework for analyzing semi-competing risks data that permits the simultaneous investigation of all three of the aforementioned scientific goals. Characterization of the induced posterior and posterior predictive distributions is achieved via an efficient Metropolis-Hastings-Green algorithm, which has been implemented in an R package. The proposed framework is applied to data on n=16,051 individuals diagnosed with pancreatic cancer between 2005-2008, obtained from Medicare Part A. We found that increased risk for readmission is associated with a high comorbidity index, a long hospital stay at initial hospitalization, non-white race, male, and discharge to home care.

October 9

Joseph J. Locascio, Ph.D.
Instructor in Neurology, Harvard Medical School
Biostatistician, Memory and Movement Disorders, Massachusetts General Hospital
Neurology and Psychiatry, Massachusetts General Hospital
Alzheimer's Disease Research Center, Massachusetts General Hospital

"Path Analysis: The Algebra of Causality"
ABSTRACT: My presentation is about "Path Analysis", also called "Structural Equation Models" (SEM) or "Causal Models". Much of my background is in applied behavioral science statistics where SEMs have been around for decades and the methodology & software for them has become fairly developed. SEMs are increasingly showing up now in the medical/biological sciences and I believe will have an increasingly prominent role in the future in cross-sectional as well as longitudinal research. I'm planning on just a basic introduction to the subject, more applied and graphically orientated rather than mathematical. My stress will be more on path analysis as a general methodological approach to causality in research, implicitly underlying many other analysis methods, rather than as just another stand-alone data analysis method itself.

October 16

Journal Club

Moderated by Rebecca Betensky, PhD; Professor, Department of Biostatistics, Harvard School of Public Health

October 30

Eyal Y. Kimchi, M.D., Ph.D.
Clinical Fellow in Neurology, Department of Neurology, Massachusetts General Hospital

"Developing a Translational Model to Determine the Pathophysiology of Delirium"
ABSTRACT: Delirium is an acute and fluctuating disturbance of attention and awareness that is most common in elderly patients. Delirium heralds the possibilities of not only sustained brain dysfunction but also dependence and death. Despite the profound and alarming nature of delirium, treatments are severely limited by an incomplete understanding of its biological basis. I am developing a translational model to determine the pathophysiology of delirium, with the hope of generating new treatments for this common and severe clinical condition. In my talk, I will provide an overview of my developing research, focusing primarily on the background and experimental design.

November 13

Journal Club

Moderated by Folefac Atem, PhD; Research Fellow, Department of Biostatistics, Harvard School of Public Health

November 20

Yared Gurmu
Doctoral Student, Department of Biostatistics, Harvard University

"Estimating the Distribution of Partnership Duration"
ABSTRACT: Data that describe sexual partnership duration are useful for modeling spread of sexually transmitted infections. Such data are commonly obtained through surveys that collect information on relationships that are ongoing during a fixed time window. This sampling mechanism leads to duration data that are left truncated and right censored; and have been analyzed using the standard truncation product limit estimator (TPLE). In this presentation, we describe a common sampling scheme for collecting sexual partnership data, discuss a key assumption required for the TPLE to be unbiased, and provide the conditions under which the nonparametric maximum likelihood estimator of the relationship duration distribution is unique and consistent. We also investigate the conditions required for the consistency of the regression coefficient from a Cox proportional hazards model that apply even when the distribution of duration is not completely identifiable due to restrictions on the support of the truncation distribution. Lastly, we will provide some illustrative examples on estimating distribution of most recent partnerships and present spline regression results based on partnership data collected from a community based AIDS prevention study in Mochudi.

December 11

Elizabeth C. Mormino, Ph.D.
Research Fellow in Neurology, Department of Neurology, Massachusetts General Hospital / Harvard Medical School

"A Status Interacts with Multiple Risk Factors to Influence Cognition in Clinically Normal Individuals"
ABSTRACT: Aberrant accumulation of beta-amyloid (A) is thought to be an early event in a biological cascade that eventually leads to Alzheimer's disease. Along these lines, many clinically normal (CN) older individuals have evidence of beta-amyloid (A) accumulation, which may be indicative of preclinical Alzheimer's disease. However, relationships between A and "downstream" brain changes within CNs are often inconsistent across research groups, suggesting that the impact of A may be modified by additional risk factors. For instance, although neurodegeneration is typically thought to occur downstream to A in models of Alzheimer's disease development, associations between A and neurodegeneration within CNs are small and there are many low A CNs with evidence of neurodegeneration. Thus, it is likely that multiple contributors of neurodegeneration exist in aging. Interestingly, CNs that have evidence of both neurodegeneration and A accumulation show the greatest risk of cognitive decline, suggesting an interaction between pathways that promote A and neurodegeneration. Additional risk factors that may influence the impact of A in aging are the presence of the APOE4 allele and female gender. Specifically, over a short term follow up, we have found that CNs that are both high A and APOE4+ show faster decline than subjects that are high A and APOE4-, suggesting that the impact of the APOE4+ allele is not purely mediated by measurable levels of A. Lastly, analyses examining the impact of gender on markers of Alzheimer's disease have revealed that although females do not harbor greater levels of A, females show worse cognition for a given level of A than males. Thus, females appear to be more vulnerable to A than their male counterparts. Overall, these analyses suggest that multiple risk factors interact with A to influence an individual's trajectory towards AD. Understanding these risk factors is essential for characterizing preclinical AD and has important implications for upcoming secondary prevention trials.

December 18

Journal Club

Moderated by Caterina Stamoulis, PhD; Assistant Professor of Radiology, Harvard Medical School; Departments of Radiology and Neurology, Children's Hospital Boston

"Characterizing global statistical significance of spatiotemporal hot spots in magnetoencephalography/ electroencephalography source space via excursion algorithms" by Yang Xu, Gustavo P. Sudre, Wei Wang, Douglas J. Weber and Robert E. Kass

January 8

Caterina Stamoulis, Ph.D.
Assistant Professor of Radiology, Harvard Medical School; Departments of Radiology and Neurology, Children's Hospital Boston

"Detection of Copy-Number Variations in Sporadic Amyotrophic Lateral Sclerosis Using the Optimized Signal Decomposition Matched-Filtering (SDMF) Method"
ABSTRACT: Amyotrophic Lateral Sclerosis (ALS) is fatal adult onset neurodegenerative disease involving gradual upper and lower motor neuron death. It is characterized by progressive muscle weakness, bulbar impairment and eventual paralysis of both limb and respiratory muscles, leading to respiratory failure and death. ALS is relatively rare, with an incidence rate of less than 1 case per 100,000. An estimated ~5,000 new cases are diagnosed every year. Mean survival time is 3-5 years. More than 90% of cases are sporadic with largely unknown genetic etiologies. Only ~10% of cases are familial and a few specific mutations have been identified, e.g. in the SOD1 or C9ORF72 genes. Copy-number variations (CNV), which are relatively large-scale (= 1000 base pairs) structural changes in the genome, have been previously identified in patients with sporadic ALS but their role in the disorder remains unclear. It has been previously suggested that multiple rare CNVs may play an important role in the pathophysiology of ALS. We have recently proposed an optimization of our previously developed Signal Decomposition Matched-Filtering (SDMF) method [Stamoulis & Betensky 2011], for robust detection of CNVs in noisy genomic data. In this talk I will discuss this optimization of SDMF and its application to genomic data from 20 patients (11 males, 9 females) with sporadic ALS, and will present preliminary findings on detected CNVs and potential implications of their locations in the genome.

January 15

Nicte Itzel Meji, M.D.
Assistant Professor, Department of Neurology, Harvard Medical School
Director, Massachusetts General Hospital Neurology Diversity and Community Outreach Initiatives

"Racial and Social Factors that Affect Parkinson's Disease Patients
ABSTRACT: Parkinson's disease (PD) affects 7-10 million people worldwide, and is expected to double in prevalence by 2030. PD patients experience decades of motor (rest tremor, rigidity, bradykinesia, gait instability) and non-motor (anxiety, depression, constipation, urinary dysfunction) symptoms as well as potential complications including falls. Both PD patients and their caregivers tend to face poor quality of life and shorter life expectancy. This talk focused on the racial and social factors that affect Parkinson's disease patients. In particular, we will: 1) discuss U.S.-wide large-database studies evaluating the effects of race, ethnicity, and social factors on PD patients, as well as 2) lay the foundation to develop targeted interventions to improve PD diagnosis, care, and outcomes for all patients. The topics at stake are of utmost importance to PD patients in an era of growing documentation of PD diagnosis, care, and clinical outcomes disparities. Discussed studies will include analyses of the Nationwide Inpatient Sample (NIS), the Medical Expenditure Panel Survey (MEPS), and the Hispanic Established Populations for the Epidemiologic Study of the Elderly (HEPESE). Proposed interventions will incorporate the role of telemedicine, including a discussion of the upcoming PCORI-funded "PD Connect" that will occur across the United States.

January 29

Folefac Atem, Ph.D.
Research Fellow, Department of Biostatistics, Harvard School of Public Health

"Estimation of the Linear Model with Right Censored Covariates"
ABSTRACT: Researchers are often faced with the problem of randomly censored covariates. The simplest and most straightforward approach for dealing with such data is to remove variables with censored observations or delete all censored observations. The former leads to model misspecification while the latter leads to overestimation of standard error due to a loss in power. In this paper we propose two approaches; the first approach is a modified Richardson and Ciampi approach based on Kaplan-Meier, the second approach is based on multiple imputation, where the multiple imputation draws are from the predictive distribution of the censored values. We used simulations to compare the performance of these approaches to that of existing methods. We apply these methods to a study of the association between beta amyloid in offspring and parental history of dementia.

February 5 (Cancelled due to inclement weather)

Jessica R. Marden
Doctoral Student, Department of Social and Behavioral Sciences, Harvard School of Public Health

"Validation of a Polygenic Risk Score for Dementia in Black and White Individuals"
ABSTRACT: Objective: To determine whether a polygenic risk score for Alzheimer's disease (AD) predicts dementia probability and memory functioning in non-Hispanic black (NHB) and non-Hispanic white (NHW) participants from a sample not used in previous genome-wide association studies.

Methods: NHW and NHB Health and Retirement Study (HRS) participants provided genetic information and either a composite memory score (n=10,401) or a dementia probability score (n=7,690). Dementia probability score was estimated for participants' age 65+ from 2006-2010, while memory score was available for participants age 50+. We calculated AD genetic risk scores (AD-GRS) based on 10 polymorphisms confirmed to predict AD, weighting alleles by beta coefficients reported in AlzGene meta-analyses. We also calculated an alternative GRS excluding APOE. We used pooled logistic regression to estimate the association of the AD-GRS with dementia probability and generalized linear models to estimate its effect on memory score.

Results: Each 0.10 unit change in the AD-GRS was associated with larger relative effects on dementia among NHW aged 65+ (OR=2.22;95% CI: 1.79,2.74; p<0.001) than NHB (OR=1.33; 95% CI: 1.00,1.77; p=0.047), although additive effect estimates were similar. Each 0.10 unit change in the AD-GRS was associated with a -0.07 (95% CI: -0.09,-0.06; p<0.001) SD difference in memory score among NHW aged 50+, but no significant differences among NHB (=-0.01; 95% CI: -0.03,0.02; p=0.546). The estimated effect of the GRS was significantly smaller among NHB than NHW (p<0.05) for both outcomes. However, a modified AD-GRS without APOE was associated with slightly larger effects on dementia probability for NHB than NHW. Conclusion: This analysis provides evidence for differential relative effects of the GRS on dementia probability and memory score among NHW and NHB in a new, national dataset.

February 12

Journal Club

Moderated by Ritesh Ramchandani; Doctoral Student, Department of Biostatistics, Harvard University

"A hybrid procedure for detecting global treatment effects in multivariate clinical trials: theory and applications to fMRI studies" by Giorgos Minas, Fabio Rigat, Thomas E. Nichols, John A.D. Aston, and Nigel Stallard

February 26

Jessica R. Marden
Doctoral Student, Department of Social and Behavioral Sciences, Harvard School of Public Health

"Validation of a Polygenic Risk Score for Dementia in Black and White Individuals"
ABSTRACT: Objective: To determine whether a polygenic risk score for Alzheimer's disease (AD) predicts dementia probability and memory functioning in non-Hispanic black (NHB) and non-Hispanic white (NHW) participants from a sample not used in previous genome-wide association studies.

Methods: NHW and NHB Health and Retirement Study (HRS) participants provided genetic information and either a composite memory score (n=10,401) or a dementia probability score (n=7,690). Dementia probability score was estimated for participants' age 65+ from 2006-2010, while memory score was available for participants age 50+. We calculated AD genetic risk scores (AD-GRS) based on 10 polymorphisms confirmed to predict AD, weighting alleles by beta coefficients reported in AlzGene meta-analyses. We also calculated an alternative GRS excluding APOE. We used pooled logistic regression to estimate the association of the AD-GRS with dementia probability and generalized linear models to estimate its effect on memory score.

Results: Each 0.10 unit change in the AD-GRS was associated with larger relative effects on dementia among NHW aged 65+ (OR=2.22;95% CI: 1.79,2.74; p<0.001) than NHB (OR=1.33; 95% CI: 1.00,1.77; p=0.047), although additive effect estimates were similar. Each 0.10 unit change in the AD-GRS was associated with a -0.07 (95% CI: -0.09,-0.06; p<0.001) SD difference in memory score among NHW aged 50+, but no significant differences among NHB (=-0.01; 95% CI: -0.03,0.02; p=0.546). The estimated effect of the GRS was significantly smaller among NHB than NHW (p<0.05) for both outcomes. However, a modified AD-GRS without APOE was associated with slightly larger effects on dementia probability for NHB than NHW. Conclusion: This analysis provides evidence for differential relative effects of the GRS on dementia probability and memory score among NHW and NHB in a new, national dataset.

March 5

Journal Club

Moderated by Jessica R. Marden; Doctoral Student, Department of Social and Behavioral Sciences, Harvard School of Public Health

March 12

Deborah Blacker, M.D., Sc.D.
Director, Massachusetts Gerontology Research Unit
Co-Director, Clinical Core, Massachusetts Alzheimer's Disease Research Center
Professor of Epidemiology, Harvard School of Public Health
Professor of Psychiatry, Harvard Medical School

"Ethical Issues in Neurological Research"
ABSTRACT: Ethical issues related to neurological research will be discussed with a focus on two key issues 1. Consent to research in the setting of impaired decision making. 2. Challenges of revealing sensitive information (e.g., PSEN1 or APOE genotype, amyloid PET scan findings).

March 26

Ritesh Ramchandani
Doctoral Student, Department of Biostatistics, Harvard University

"A Generalized Global Rank Test for Multiple, Possibly Censored, Outcomes"
ABSTRACT: In clinical studies, a single outcome does not always adequately capture the effect of an intervention, so other outcomes are often considered as well. The design, analysis, and interpretation of studies in the presence of multiple outcomes like these can be difficult. We propose a general nonparametric scoring method for combining two or more endpoints into a single summary statistic for efficacy. This general test encompasses other global or composite rank tests that have been previously proposed, and can accommodate censored observations. First we score each pair of subjects with respect to each outcome, and then reduce the dimension of the multiple pairwise scores to get a composite score for the pair of subjects. A rank-sum type test on the composite scores is then performed. The method of dimension reduction is investigator chosen, and we propose methods for choosing outcome weights to improve the power of certain tests under specific alternative hypotheses. Examples of various tests are provided using an ALS dataset.

April 2

Paola Gilsanz, MPH
Doctoral Candidate, Department of Social and Behavioral Sciences, Harvard School of Public Health

"Short- and Long-Term Depressive Symptoms and Arrhythmic Pathways to Stroke"
ABSTRACT: Emerging data suggests that elevated depressive symptoms predict stroke onset. However, it remains unclear whether this relationship is causal. Furthermore, despite substantial research, we have not been able to establish the mechanisms linking depression and stroke. We used data from two complementary longitudinal studies, the Health and Retirement Study (HRS) and the Cardiovascular Health Study (CHS), to examine two aims. The first aim of this was to determine what how changes in depressive symptoms across two consecutive assessments impact first incidence of all stroke types among middle aged and elderly individuals in the United States. Our second aim was to examine atrial fibrillation as a partial mediator of the association between depressive symptoms and onset of ischemic stroke. Understanding how changes in depressive symptoms predict stroke and the underlying etiology of the relationship can help health practitioners and identify patients at greater risk for stroke. Identifying whether atrial fibrillation is a mediating factor in the relationship between depression and stroke is clinically important as it may inform treatment plans of at risk patients.

April 9

Journal Club

Moderated by Eyal Y. Kimchi, M.D., Ph.D.; Clinical Fellow in Neurology, Department of Neurology, Massachusetts General Hospital

"Dynamic Analysis of Learning in Behavioral Experiments" by Anne C. Smith, Loren M. Frank, Sylvia Wirth, Marianna Yanike, Dan Hu, Yasuo Kubota, Ann M. Graybiel, Wendy A. Suzuki, and Emery N. Brown

April 23

Brian C. Healy, Ph.D.
Assistant Professor of Neurology, Department of Neurology, Brigham and Women's Hospital

"Modeling Disease Progression in Patients with Multiple Sclerosis"
ABSTRACT: Multiple sclerosis (MS) is the most common neurologic disease among young people in the US. The majority of MS patients initially have a relapsing-remitting form of the disease, and most of the available treatments have been approved by the FDA by demonstrating an effect in terms of reducing the number of relapses. Unfortunately, the more devastating part of the disease is neurodegeneration, and the impact of the available treatments on neurodegeneration is inconclusive. One of the problems with estimating the treatment effect on the neurodegenerative component of the disease is related to challenges in directly measuring neurodegeneration. The primary clinical outcome measure in most clinical trials is the presence of sustained progression on the EDSS. Using data from the Comprehensive Longitudinal Investigation of Multiple Sclerosis at the Brigham and Women's Hospital (CLIMB), sample size calculations for analyses using sustained progression as the outcome were compared to sample size calculations for alternative analysis strategies. Furthermore, additional problems with the definition of sustained progression were observed when longer term follow-up was considered. Given the problems associated with sustained progression as an outcome, two alternative analysis approaches are proposed. The first is a joint first order Markov model for the changes in the EDSS and the presence/absence of a relapse, and the second is a rescaling of the EDSS to allow direct comparison of the observed changes in the EDSS.

April 30

Kush Kapur, Ph.D.
Instructor in Neurology, Department of Neurology, Children's Hospital Boston

"Sample Size Determination for Longitudinal Designs with Binary Response"
ABSTRACT: In this talk I will discuss appropriate statistical methods for determining the required sample size while comparing the efficacy of an intervention to a control with repeated binary response outcomes. Our proposed methodology incorporates the complexity of the hierarchical nature of underlying designs and provides solutions when varying attrition rates are present over time. We explore how the between subject variability and attrition rates jointly influence the sample size formula. A practical guideline is provided when information regarding individual variance components is unavailable. The validity of our methods is established by extensive simulation studies. Results are illustrated with the help of two randomized clinical trials in the areas of contraception and insomnia.

May 7

Journal Club

Moderated by Nicte Itzel Meji, M.D.; Assistant Professor, Department of Neurology, Harvard Medical School
Director, Massachusetts General Hospital Neurology Diversity and Community Outreach Initiatives


"Incidence of Parkinson's Disease: Variation by Age, Gender, and Race/Ethnicity" by Stephen K. Van Den Eeden, Caroline M. Tanner, Allan L. Bernstein, Robin D. Fross, Amethyst Leimpeter, Daniel A. Bloch, and Lorene M. Nelson and "The Frequency of Idiopathic Parkinson's Disease by Age, Ethnic Group, and Sex in Northern Manhattan, 1988-1993" by Richard Mayeux, Karen Marder, Lucien J. Cote, Jean Denaro, Nancy Hemenegildo, Helen Mejia, Ming-Xin Tang, Rafael Lantigua, David Wilder, Barry Gurland, and Allen Hauser

May 14

Lorenzo Trippa, Ph.D.
Assistant Professor, Department of Biostatistics, Harvard School of Public Health and Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute

"Bayesian Nonparametric Cross-Study Validation of Prediction Methods"
ABSTRACT: We consider comparisons of statistical learning algorithms using multiple datasets, via leave-one-in cross-study validation: each of the algorithms is trained on one dataset; the resulting model is then validated on each remaining dataset. This poses two statistical challenges that need to be addressed simultaneously. The first is the assessment of study heterogeneity, with the aim of identifying subset of studies within which algorithm comparisons can be reliably carried out. The second is the comparison of algorithms using the ensemble of datasets. We address both problems by integrating clustering and model comparison. We formulate a Bayesian model for the array of cross-study validation statistics, which defines clusters of studies with similar properties, and provides the basis for meaningful algorithm comparison in the presence of study heterogeneity. We illustrate our approach through simulations involving studies with varying severity of systematic errors, and in the context of medical prognosis for patients diagnosed with cancer, using high-throughput measurements of the transcriptional activity of the tumor's genes.

May 21

Journal Club

Moderated by Josephine Asafu-Adjei, PhD; Research Fellow, Department of Biostatistics, Harvard School of Public Health

"Survival Model Predictive Accuracy and ROC Curves" by Patrick J. Heagerty and Yingye Zheng



Back to HSPH Biostatistics Maintained by the Biostatistics Webmaster
Last Update: May 12, 2014