Estimating Brain Connectivity Networks With Additional Knowledge
This repository provides the data and an implementation of the algorithms described in the paper Integrating Additional Knowledge Into the Estimation of Graphical Models for Brain Connectivity Networks by Yunqi Bu and Johannes Lederer.
Implementation/graph_estimation.R contains a function
GraphEstimation to estimate
a connectivity network using the
SI method described in the paper.
In order to use this function, one needs a matrix of fMRI data and a matrix containing pairwise distances between the brain regions.
An exemplary matrix of fMRI data (preprocessed data for one subject) and the distance matrix for the regions can be found in this
A program that uses these matrices and produces an image of the resulting
connectivity network can be found in
Detailed descriptions on the fMRI preprocessing procedure can be found under
Preprocessed data for all 37 patients in the study can be found under
Data/PreprocessedData. The brain region names, the lobe that each region belongs to, and the 42 regions of interest indicator is given in
The coding invokes Google's style guide for
- Yunqi Bu — Graduate student in Biostatistics, University of Washington — methodology and
- Johannes Lederer — Assistant Professor in Statistics, University of Washington — methodology and
- Benjamin J. Phillips — Undergraduate student in Mathematics, University of Washington — revision of
The raw fMRI data was collected and provided by Dr. Dantao Peng, Dr. Yanlei Mu, and Dr. Xiao Zhang. The data preprocessing was conducted by Dr. Min Zhang.
Cite as "Bu and Lederer, Integrating Additional Knowledge Into the Estimation of Graphical Models for Brain Connectivity Networks, arXiv:1704.02739, 2017"