A Bayesian Framework for Joint Analysis of Heterogeneous Neuroscience Data

Project Summary

This paper addresses analysis of heterogeneous data, such as ordered, categorical, real and count data. Such data are of interest in our motivating application, cognitive and brain science, in which subjects may answer questionnaires, and also (separately) undergo fMRI interrogation. A contribution of this paper concerns the joint analysis of how people answer questionnaires and how their brain responds to external stimuli (here visual), the latter measured via fMRI.

Themes and Categories

In this paper we ask a novel and practical question, which to our knowledge has not been considered previously: can one predict the fMRI response (here from the amygdala and the ventral striatum) to external stimuli, based upon knowledge of how the subject answers a questionnaire and genetic data?

A new model is developed for joint analysis of ordered, categorical, real and count data. In the motivating application, the ordered and categorical data are answers to questionnaires, the (word) count data correspond to the text questions from the questionnaires, and the real data correspond to fMRI responses for each subject. We also combine the analysis of these data with single-nucleotide polymorphism (SNP) data from each individual. The questionnaires considered here correspond to standard psychological surveys, and the study is motivated by psychology and neuroscience. The proposed Bayesian model infers sparse graphical models (networks) jointly across people, questions, fMRI stimuli and brain activity, integrated within a new matrix factorization based on latent binary features. We demonstrate how the learned model may take fMRI and SNP data from a subject as inputs, and predict (impute) how the individual would answer a psychological questionnaire; going in the other direction, we also use an individual's SNP data and answers from questionnaires to impute unobserved fMRI data. Each of these two imputation settings has practical and theoretical applications for understanding human behavior and mental health, which are discussed.





Related People

Related Projects

A team of students led by faculty and researchers from the School of Medicine, the Center for Global Reproductive Health at the Duke Global Health Institute, and the Duke Evidence Lab will collaborate on the user interface for a tool developed to help advocates and policymakers target family planning resources to key populations in low resource populations. Team members will traverse the app development lifecycle while contributing to a tool that can improve global reproductive health.

Faculty Lead: Megan Huchko

Project Manager: Amy Finnegan

A team of students led by Drs. Aquino (Engineering) and Routh (Urologic Surgery) will develop objective algorithms in order to guide data interpretation from a urology test, known as urodynamics, which is used in children with spina bifida in order to define a patient’s risk of debilitating bladder and kidney complications.  Urodynamics involves dynamic pressure monitoring as the bladder is filled with fluid.  This project is part of a 21-institution collaboration coordinated and funded by the U.S. Centers for Disease Control and Prevention (CDC), with the long-term goal of defining optimal management strategies for children with spina bifida. The short-term goal of this Data+ application is to define initial features of urodynamics that can be applied to increasingly complex future algorithms in order to guide clinical interpretations that determine whether, for example, children need reconstructive surgery in order to avoid complications of their disease.

Faculty Leads: Wilkins Aquino, Jonathan Routh

A team of students led by researchers from the Michael W. Krzyzewski Human Performance Laboratory (K-Lab) will develop an analytic and report generating web-based application to help the K-Lab reduce musculoskeletal injuries in student-athletes at Duke University. This tool will produce actionable, student-athlete-specific reports that incorporate the analysis of previous injury history and current capabilities (K-Lab assessments) in order to identify injury risk and develop individualized recommendations for injury prevention. Students will develop analytic tools and scoring criteria to assess injury risk through profiling of data based on minimally clinically important differences, injury profiles, peer group analysis, and injury risk scoring strategies based on a comprehensive set of performance metrics. Injury risk identification will be furthered enhanced by clustering data analysis around joint or tissue specific injury risk, previous injury history, and athlete capabilities (strength, flexibility, and postural stability). The final deliverable will enhance injury prevention strategies for student-athletes and other populations by bridging the analytic gap between injury risk screening and actionable injury prevention strategies.

Faculty Lead: Dr. Tim Sell

Project Manager: Brinnae Bent