A Sparse and Spatially Constrained Generative Regression Model for fMRI Data Analysis

• Published in: IEEE transactions on biomedical engineering Vol. 59; no. 1; pp. 58 - 67
• Main Authors: Oikonomou, Vangelis P., Blekas, Konstantinos, Astrakas, Loukas
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.01.2012; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Analytical models; Biological and medical sciences; Brain - physiology; Computer Simulation; Computerized, statistical medical data processing and models in biomedicine; Correlation; Data Interpretation, Statistical; Data models; Estimation; Expectation maximization (EM) algorithm; functional magnetic resonance imaging (fMRI) analysis; general linear regression model (GLM); Humans; Image Interpretation, Computer-Assisted - methods; Magnetic Resonance Imaging - methods; Markov processes; Markov random field (MRF); Mathematical model; Medical management aid. Diagnosis aid; Medical sciences; Models, Neurological; Models, Statistical; Nerve Net - physiology; Noise; Regression Analysis; relevance vector machine (RVM); functional magnetic resonance imaging (fMRI) analysis; Data analysis; Markov random field (MRF); Nuclear magnetic resonance imaging; Linear model; Random field; general linear regression model (GLM); relevance vector machine (RVM); Relevance criterion; Expectation maximization (EM) algorithm; Medical imagery; EM algorithm; Biomedical engineering; Functional imaging
• ISSN: 0018-9294; 1558-2531
• DOI: 10.1109/TBME.2010.2104321

In this study, we present an advanced Bayesian framework for the analysis of functional magnetic resonance imaging (fMRI) data that simultaneously employs both spatial and sparse properties. The basic building block of our method is the general linear regression model that constitutes a well-known probabilistic approach. By treating regression coefficients as random variables, we can apply an enhanced Gibbs distribution function that captures spatial constrains and at the same time allows sparse representation of fMRI time series. The proposed scheme is described as a maximum a posteriori approach, where the known expectation maximization algorithm is applied offering closed-form update equations for the model parameters. We have demonstrated that our method produces improved performance and functional activation detection capabilities in both simulated data and real applications.