CMStatistics 2015: Start Registration
View Submission - CMStatistics
B0342
Title: A functional data approach for using EEG Data to select treatment for major depressive disorder Authors:  Adam Ciarleglio - New York University School of Medicine (United States) [presenting]
Eva Petkova - New York University (United States)
Thaddeus Tarpey - Wright State University (United States)
Todd Ogden - Columbia University (United States)
Abstract: Major depressive disorder (MDD) is a disease characterized by substantial heterogeneity. This heterogeneity exists in both the symptoms associated with MDD and in the responses that MDD patients have to various forms of treatment: what works for one patient may be ineffective or harmful for another. This makes treatment selection a difficult task particularly because there are no widely accepted biomarkers for MDD treatment response. Recently, the search for such biomarkers has broadened to include measures derived from neuroimaging modalities such as MRI, fMRI, and EEG that are collected at baseline. This seems justified since various aspects of brain structure and function have been implicated in depressive symptoms and in response to treatment. We propose a functional data approach for using baseline EEG data to both select the best MDD treatment for the individual and provide interpretable measures of the relationship between the EEG signal and treatment response. The approach will be evaluated in several realistic settings using synthetic data and will be applied to real data arising from a multi-center clinical trial comparing two treatments for MDD in which subjects had their EEG data collected at baseline.