A Global Bayes Factor for Observations on an Infinite-Dimensional Hilbert Space, Applied to Signal Detection in fMRI

Functional Magnetic Resonance Imaging (fMRI) is a fundamental tool in advancing our understanding of the brain's functionality. Recently, a series of Bayesian approaches have been suggested to test for the voxel activation in different brain regions. In this paper, we propose a novel definition for the global Bayes factor to test for activation using the Radon-Nikodym derivative. Our proposed method extends the definition of Bayes factor to an infinite dimensional Hilbert space. Using this extended definition, a Bayesian testing procedure is introduced for signal detection in noisy images when both signal and noise are considered as an element of an infinite dimensional Hilbert space. This new approach is illustrated through a real data analysis to find activated areas of Brain in an fMRI data.

[1]  R. Turner,et al.  Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. , 1992, Proceedings of the National Academy of Sciences of the United States of America.

[2]  Tom Fleischer,et al.  Applied Functional Analysis , 2016 .

[3]  N. Krylov,et al.  Introduction to the Theory of Random Processes , 2002 .

[4]  L. M. M.-T..  Theory of Probability , 1929, Nature.

[5]  Alan C. Evans,et al.  A Three-Dimensional Statistical Analysis for CBF Activation Studies in Human Brain , 1992, Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism.

[6]  K. Worsley.  Testing for signals with unknown location and scale in a χ2 random field, with an application to fMRI , 2001, Advances in Applied Probability.

[7]  José M. Bernardo,et al.  Bayesian Statistics , 2011, International Encyclopedia of Statistical Science.

[8]  Karl J. Friston,et al.  Mixtures of general linear models for functional neuroimaging , 2003, IEEE Transactions on Medical Imaging.

[9]  D. Siegmund,et al.  Testing for a Signal with Unknown Location and Scale in a Stationary Gaussian Random Field , 1995 .

[10]  H. Vincent Poor,et al.  An Introduction to Signal Detection and Estimation , 1994, Springer Texts in Electrical Engineering.

[11]  Karl J. Friston,et al.  Statistical parametric maps in functional imaging: A general linear approach , 1994 .

[12]  Sophie Achard.  The Statistical Analysis of Functional MRI Data , 2008 .

[13]  H. Kuo.  Gaussian Measures in Banach Spaces , 1975 .

[14]  Hal S. Stern,et al.  A Nonparametric Bayesian Approach to Detecting Spatial Activation Patterns in fMRI Data , 2006, MICCAI.

[15]  Keith J. Worsley,et al.  Rotation space random fields with an application to fMRI data , 2003 .

[16]  Bertrand Thirion,et al.  A fully Bayesian approach to the parcel-based detection-estimation of brain activity in fMRI , 2008, NeuroImage.

[17]  R. Adler,et al.  Random Fields and Geometry , 2007 .

[18]  Balram S. Rajput,et al.  Equivalent Gaussian measure whose R-N derivative is the exponential of a diagonal form , 1981 .

[19]  William D. Penny,et al.  Bayesian fMRI data analysis with sparse spatial basis function priors , 2007, NeuroImage.

[20]  M. R. Farid Rohani,et al.  A bayesian signal detection procedure for scale‐space random fields , 2006 .

[21]  Karl J. Friston,et al.  Comparing Functional (PET) Images: The Assessment of Significant Change , 1991, Journal of cerebral blood flow and metabolism : official journal of the International Society of Cerebral Blood Flow and Metabolism.

[22]  Dale E. Varberg.  Equivalent Gaussian Measures with a Particularly Simple Radon-Nikodym Derivative , 1967 .