Agreement Methods for Complex Outcomes in Biomedical Studies Público
Dai, Tian (2016)
Published
Abstract
In biomedical studies, researchers are often interested in evaluating the similarity of measurements produced by different methods on the same subjects. Many methods have been proposed for measuring agreement between standard outcomes. Limited work has been done to assess agreement for complex biomedical outcomes. In this dissertation, we propose novel agreement methods for two complex outcomes often encountered in biomedical studies, survival data and high-dimensional neuroimaging data. First, we study the local agreement pattern between two survival outcomes. Most classical agreement methods have been focused on global summary statistics, which cannot be used to describe various local agreement patterns. In this work, we propose a new agreement measure based on bivariate hazard functions to characterize the local agreement pattern between two correlated survival outcomes. The proposed measure naturally accommodates censored observations, fully captures the dependence structure between bivariate survival times and provides detailed information on how the strength of agreement evolves over time. Next, we investigate statistical methods for assessing reproducibility of imaging data in multi-site studies. Considering the special features of imaging data, such as high dimensionality, we propose a two-stage network-based method to effectively assess the similarity between the same subject's brain images acquired at different sites. In the first stage, we reduce the dimensionality of imaging data by extracting active functional networks under experimental conditions and estimating the corresponding temporal dynamics. In the second stage, we propose functional agreement indices to measure the agreement between the same subject's network-specific temporal dynamics estimated across different sites. Last, we propose a prediction method based on Bayesian hierarchical model that uses individual's earlier scans, coupled with relevant baseline characteristics, to predict the individual's future functional connectivity. The proposed method could provide a useful tool to predict the changes in individual patient's brain connectivity with the progression of disease. It can also be used to predict a patient's brain connectivity after a specified treatment regimen which could potentially help guide individualized treatment plan. Another utility of the proposed method is that it could be applied to test-retest reproducibility imaging data to develop a more reliable estimator for individual functional connectivity.
Table of Contents
1 Overview. 1
1.1 Motivation. 3
1.2 Outline. 6
2 Agreement Methods for Survival Outcomes. 9
2.1 Introduction. 11
2.2 Methods. 14
2.2.1 Proposed local agreement pattern measure φ(t1,t2). 14
2.2.2 Properties of φ(t1,t2) as a local agreement pattern measure. 16
2.2.3 Estimation and inference for φ(t1,t2). 19
2.2.4 Choice of the kernel functions and bandwidths for estimating the hazard functions. 22
2.3 Simulation Studies. 24
2.4 Data Example. 26
2.5 Remarks. 29
2.6 Appendix. 30
3 Agreement Methods for High Dimensional Neuroimaging Data. 40
3.1 Introduction. 42
3.2 Method. 46
3.2.1 Two-stage agreement methods for task-fMRI data. 46
3.2.2 Estimation and Inference. 50
3.3 fBIRN Data Example. 53
3.3.1 Preprocessing procedure. 53
3.3.2 First Stage. 53
3.3.3 Second Stage. 54
3.4 Simulation Studies. 61
3.5 Remarks. 64
3.6 Appendix. 64
4 Predicting Brain Functional Connectivity in Resting-state fMRI Data Using a Bayesian Hierarchical Model. 72
4.1 Introduction. 74
4.2 Method. 76
4.2.1 Bayesian Hierarchical Model(BHM). 77
4.2.2 Prediction Algorithm. 79
4.2.3 Model Specification for Longitudinal Studies. 82
4.2.4 Model Specification for Test-retest Studies. 83
4.2.5 Connections and Differences with a Shrinkage Estimator in Test-retest Studies. 85
4.3 Simulation Studies. 87
4.4 Data Applications. 94
4.4.1 Application to a Longitudinal ADNI2 Study. 94
4.4.2 Application to a Test-retest Kirby21 Study. 102
4.5 Remarks. 103
4.6 Appendix. 106
5 Summary and Future Work. 108
Bibliography. 113
About this Dissertation
School | |
---|---|
Department | |
Degree | |
Submission | |
Language |
|
Research Field | |
Palabra Clave | |
Committee Chair / Thesis Advisor | |
Committee Members |
Primary PDF
Thumbnail | Title | Date Uploaded | Actions |
---|---|---|---|
|
Agreement Methods for Complex Outcomes in Biomedical Studies () | 2018-08-28 11:44:15 -0400 |
|
Supplemental Files
Thumbnail | Title | Date Uploaded | Actions |
---|---|---|---|
|
paper2_simufig1.jpg () | 2018-08-28 11:47:29 -0400 |
|
|
IC3VN1.pdf () | 2018-08-28 11:49:03 -0400 |
|
|
simu_plot11.pdf () | 2018-08-28 11:51:06 -0400 |
|
|
simu_plot12.pdf () | 2018-08-28 11:53:12 -0400 |
|
|
inconf_ADNI2.png () | 2018-08-28 11:54:31 -0400 |
|
|
IC1MN.pdf () | 2018-08-28 11:55:58 -0400 |
|
|
MSE_Kirby21.png () | 2018-08-28 11:57:27 -0400 |
|
|
simu_plot2.pdf () | 2018-08-28 11:59:02 -0400 |
|
|
simu_plot1.pdf () | 2018-08-28 12:00:47 -0400 |
|
|
post_Delta_abs02_cut07.pdf () | 2018-08-28 12:02:18 -0400 |
|
|
dataplot3.pdf () | 2018-08-28 12:04:51 -0400 |
|
|
figure2_1.pdf () | 2018-08-28 12:06:56 -0400 |
|
|
ICmaps.pdf () | 2018-08-28 12:08:29 -0400 |
|
|
dataplot2.pdf () | 2018-08-28 12:09:55 -0400 |
|
|
baseline_Contrast_5.png () | 2018-08-28 12:11:19 -0400 |
|
|
data1_plot1015.png () | 2018-08-28 12:13:04 -0400 |
|
|
simu_plot22.pdf () | 2018-08-28 12:14:31 -0400 |
|
|
ts-IC19.pdf () | 2018-08-28 12:16:18 -0400 |
|
|
ts-IC14.pdf () | 2018-08-28 12:18:41 -0400 |
|
|
heterogen.jpg () | 2018-08-28 12:20:05 -0400 |
|
|
change_summary_threshPP_AD.jpg () | 2018-08-28 12:21:40 -0400 |
|
|
simu_longitudinal_3methods.pdf () | 2018-08-28 12:23:18 -0400 |
|
|
simu_replicated_fig2.pdf () | 2018-08-28 12:24:38 -0400 |
|
|
simu_plot21.pdf () | 2018-08-28 12:25:49 -0400 |
|
|
simu_longitudinal_fig2_3methods.pdf () | 2018-08-28 12:26:55 -0400 |
|
|
inconf_Kirby21.png () | 2018-08-28 12:27:59 -0400 |
|
|
homogen.jpg () | 2018-08-28 12:29:07 -0400 |
|
|
baseline_corr.png () | 2018-08-28 12:31:18 -0400 |
|
|
MSE_ADNI2.png () | 2018-08-28 12:33:03 -0400 |
|
|
ADNI2_MSE_individual.pdf () | 2018-08-28 12:34:44 -0400 |
|
|
simu_replicated_1.pdf () | 2018-08-28 12:37:14 -0400 |
|