I'm primarily interested in applying novel statistical techniques to the analysis of data arising from neuroimaging studies and other sources of large observational data. My work fuses ideas from causal inference with machine learning models to reduce the effects of confounding on estimated neurological disease patterns. I'm interested in developing similar methods to study mediation and to learn about covariance patterns among multiple imaging modalities. My work is motivated by the need to understand and quantify how neurological diseases cause deviations from the structural and functional characteristics of healthy brains. I have applied statistical methods to neuroimaging data to study multiple sclerosis, Alzheimer's disease, epilepsy, and adolescent development.

Some of the images below link to more detailed descriptions of my research. I'm in the process of adding more content, so please check back soon for updates.

SVM feature standardization for multivariate pattern analysis

Addressing confounding in multivariate pattern analysis

Data fusion techniques for multimodal neuroimaging studies

Optimal dynamic treatment regimes for probabilities and quantiles