Dr. Heather Shappell

Dr. Heather Shappell Uses Data Set Analysis to Advance Brain Science

Heather Shappell
Lecturer, Computer Science
Postdoctoral Fellow, Johns Hopkins University, Bloomberg School of Public Health

PhD, MA, Boston University; BS, Arcadia University

What is your area of expertise?
I am biostatistician, which means I analyze data related to public health and medical studies. My specialty is in the statistical analysis of network data, with a particular focus on applications to neuroscience and brain networks. I have also been involved in statistical analysis for several clinical trials, including clinical trials for the treatment of the rare disease progeria, as well as in the analyses for observational studies studying cardiovascular disease and mental illness.

Please tell us about your work. Can you share any current research or recent publications?
My main research interest is in developing statistical methods to estimate and study functional brain networks from fMRI (functional magnetic resonance imaging) data. An fMRI scan is used to measure brain activity by detecting changes associated with blood flow. From this data, we seek to estimate a network where the nodes represent brain regions and the edges represent synchrony or coupling between pairs of brain regions. In other words, two brain regions with an edge between them would represent two regions that may be communicating or working together.

In my area of work, I develop methods to estimate one of these brain networks at every point of the fMRI brain scan for a particular subject (this is called dynamic functional connectivity). I also develop methods to perform statistical tests that can be used to compare these networks across people. In a recent publication of mine relating to this work, “Improved state change estimation in dynamic functional connectivity using hidden semi-Markov models,” my collaborators and I suggested a new modeling framework for estimating brain networks and comparing them across individuals.

How does the subject you work in apply in practice? What is its application?
More and more, dynamic functional connectivity analysis is being used to help understand how the brain works. It is also being used to better understand the mechanisms behind psychological disorders and conditions. A recent study I have been working on with collaborators from Johns Hopkins Medical School and UNC Chapel Hill, for example, involves children with attention-deficit/hyperactivity disorder (ADHD). After performing a dynamic connectivity analysis, our main findings indicate that children with ADHD spend less total time in, and switch more quickly out of, an anti-correlated brain state, characterized by a greater number of negative correlations between the default mode network and other sub-networks; particularly the salience/ventral attention network. Additionally, we found that children with ADHD spent more time in a hyper-connected state. This knowledge may lead to better treatment for children with this condition.

What courses do you teach at MET?
I teach Foundations of Analytics with R (MET CS 544) and Data Analysis and Visualization (MET CS 555).

Please highlight a particular project within these courses that most interests your students. What “real-life” exercises do you bring to class?
Both courses I teach incorporate a final project. Students find a data set that interests them, sometimes drawn from their professional career, and apply the statistical techniques taught throughout the course. It’s a fun opportunity to be creative, and we generally receive positive feedback on this aspect of the course. Students are able to think about the skills they have learned in a setting that interests them, sometimes to the direct benefit of their career.

Based on your background and experience, what do you consider to be the unique value you bring to the classroom?
I have quite a few years of consulting experience, working with researchers from a variety of medical and public health fields. I like to bring this perspective to the classroom. Rather than only teaching the statistical theory and methods (or the programming skills), I am able to discuss examples of real studies I have worked on. It brings the methods to life. Students like to know that these methods are actually applied in the real world, and they benefit from and enjoy hearing exactly how they are applied.

I am very passionate about teaching, and I love to meet and work with students! I encourage anyone interested in data science, numbers, analytics, etc. to sign up for Foundations of Analytics with R (MET CS 544) and Data Analysis and Visualization (MET CS 555)!

View all posts