Dear SPM users
If you like to reviewer the following paper submitted to IEEE Transactions
on Biomedical Engineering, please send me your postal and email address and
phone and fax numbers.
Thank you for your kind consideration
Jagath Rajapakse, Ph.D.
Assitant Professor
Nanyang Technological University, Singapore.
Email: [log in to unmask]
----------------------------------------------------------------------------
-------------------------
TITLE: Bayesian Approach to Segmenataion of Statistical Maps
ABSTRACT
A contextual segmentation technique to detect brain activation
from functional brain images is presented in the
Bayesian framework. Unlike previous similar approaches
a Markov random field (MRF) is used to represent activation
configuration of the brain voxels and the likelihoods given by
statistical parameter maps (SPMs) are directly used
to find the maximum a posteriori (MAP) estimation of segmentation. Further
the approach incorporates the hemodynamic modulation function for the
analysis
and can analyze experiments involving multiple input stimuli.
A segmentation algorithm using a simulated annealing scheme
is proposed and simulation results and comparisons with the simple
thresholding and the statistical parameter mapping (SPM)
approaches are presented with synthetic images and the functional MR images
acquired in visual and memory tasks and event-related working memory tasks.
Experiments show that an MRF is a
valid representation for the activation obtained in functional brain
images and the present technique renders a superior segmentation scheme
to the context-free approach and the dual-thresholding technique in the
SPM approach.
{\em Key words:}
Bayesian methods,
functional brain imaging,
functional MRI,
Gaussian random fields,
image segmentation,
statistical parameter mapping,
Markov random fields,
MAP estimation.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|