Research.We use machine learning techniques for analysis of neuroimaging and other types of data on mental illness, to aid diagnosis and treatment response prognosis.
|
DatasetsIn our research, we use large imaging datasets on a variety of brain disorders, including ADHD, schizophrenia, depression, Alzheimers, and OCD. We also work with other types of data including clinical trials, EEG, eye-tracking, clinical and neurocognitive, and genetics data.
|
TechniquesMachine LearningWe use a variety of machine learning methods for predictive analysis of mental illness data. These include "off-the-shelf" classifiers such as support vector machine (SVM), Logistic Regression, Decision Trees, Naïve Bayes, and Nearest Neighbors, as well as more novel approaches such as probabilistic graphical models (PGM) and Deep Learning.
ImagingWe use a variety of imaging modalities including structural MRI, functional MRI, and DTI. Depending on the research question, we may extract various features from imaging data, including but not limited to functional connectivity or various other brain network features. (see our Publications for a list of Papers and the modalities used in each of them.)
|