Introduction

BAX is a Matlab-based analysis tool for functional MRI datasets. It can be used for incremental or sliding window GLM-based analysis, correlation analysis, or consistency map analysis. The current version is 0.3.

BAX runs on Matlab (MathWorks, Inc.) either in Linux or Windows operating systems. It also requires some routines from SPM2 (Wellcome Department of Cognitive Neurology, London, UK) for its operation. SPM2 can be freely downloaded from this website: http://www.fil.ion.ucl.ac.uk/spm/.

Installation

BAX can be downloaded here. To install BAX, simply uncompress the package into a directory in your local machine. You can then run Matlab and include the directory in your path. Be sure to include SPM also in your path. Type baxgui in Matlab's prompt to run BAX.

Changes

BAX Version 0.3
6 Sep 2007

1. Correlation analysis is now supported. This can be found in the CMap menu with two additional submenus: Generate correlation map and Display correlation map. In the Generate correlation map, you can also choose whether to use the t-statistics series (From t-statistics series) or the original bold series (From bold series). To use this one, you need to define first a region of interest (ROI). The series of this ROI is then average and is correlated to the series of the entire brain. The output is stored in a file called corr_ROI-name_stat_inc.mat, for t-series generated using incremental analysis, corr_ROI-name_stat_swa.mat for t-series generated using sliding window analysis, or corr_ROI-name_bold.mat for bold series. Use the analysis mode frame to specify which t-series to use in the correlation analysis.
2. The output files (corr_*.mat) contains the following variables: corr for the estimated correlation coefficient, df for the degrees of freedom, index for indices of the chosen ROI, and refsig for the reference signal, which is the mean of the ROI's time series.
3. The display correlation map submenu can be used if you want to display an existing correlation map generated in an earlier analysis. You can choose which file to display. The application will read the file, estimate the t-statistics value for each computed correlation, compute the probability of the estimated t-value, and use the specified p-value to have a threshold. Voxels with significant t-value will be displayed in red.
4. The correlation coefficients are estimated using bax_corr_inc function (in bax_corr_inc.m file), which implement an alternative incremental approach to estimate the correlation coefficient. If so desired, you can switch to the standard algorithm using bax_corr (in bax_corr.m file)
5. bax_notes.m describes some information for the different file formats needed by baxgui, in particular, the format to specify ROI files from an MNI coordinates, center coordinates, or saved ROIs, among other information. Example files are also included (mnicoord.txt, centercoord.txt and ROImatfile.mat)
6. The about box is now available. This is accessible from the Help menu.

This note is also available in the README.txt file included in the package.