Independent Component Analysis of fMRI Data with Application to the Study of Functional Connectivity in Neurodegeneration

• Activity Number: 565
• Type: Topic Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Dai Feng*+ and Richard Baumgartner and Alexandre Coimbra and Marie Holahan and Jacquelynn Cook and Donald Williams
• Companies: Merck & Co., Inc. and Merck Research Laboratories and Merck & Co., Inc. and Merck & Co., Inc. and Merck & Co., Inc. and Merck & Co., Inc.
• Keywords: independent component analysis ; Bayesian methods ; fMRI ; functional connectivity ; neurodegeneration

Abstract

Independent component analysis approach (ICA) is powerful due to its potential capabilities to account for unknown yet structured processes, which can be found in fMRI studies. Various Bayesian methods are used in model order selection and component density estimation. Currently there is a lot of interest in biomarkers derived from the functional connectivity in fMRI. Reports of successful application include Alzheimer's disease, schizophrenia, depression, and chronic pain. In our contribution, we will discuss statistical issues in ICA as applied to the studies of fMRI connectivity of aging and neurodegeneration in non-human primates. We will show how meaningful endpoints can be derived using ICA and compare them with those obtained with traditional "seed" voxel approach used in imaging. Finally we will elaborate on the possibility of using ICA-based endpoints as biomarkers for progressi