Machine learning

Created multi-threaded parallel pipeline to analyze brain MRI images across multiple datasets (weight loss, Parkinson’s, childhood obesity); speeding up processing by 800%. Applied statistical mediation analysis to uncover causal links between fitness, aspects of brain health and cognitive function. Techniques: Matlab, FSL, Freesurfer, gnu parallel, SPSS, mediation

I used machine learning to predict cardiovascular risk factor ie. blood pressure variability based on brain data as subjects performed stressful tasks inside fMRI scanner. People with a tendency to exhibit exaggerated blood pressure reactions to psychological stressors are at risk for hypertension, adverse clinical cardiovascular events, and premature cardiovascular mortality. Exaggerated blood pressure reactions to psychological stressors may be determined, in part, by a “brain phenotype” that is characterized by reliable neural activity changes in brain areas that regulate cardiovascular physiology during stressful experiences.

Executed a complete human subject study: Designing experimental protocol including visual psychological tasks, Recruiting stroke patients with visual attention deficits and healthy control subjects Classifying brain activity using various multivariate pattern classifiers ( i.e. SVM) Statistical testing of research hypotheses using SPSS (ANOVA) Techniques : Psychtoolbox, Python, libSVM, BrainVoyager, medical record review, SPSS

Created Predictive flow algorithm: discovers the relationship between brain regions based on their temporal sequences of activation using machine learning. [Github] Analyzed one of the largest datasets for brain neuroimaging: 800+ participants, 220 temporal scans, 200K data point per scan (~ 1 TB) Techniques: Matlab, Python, R, Jupyter, L1 regularized regression (LASSO)