Martinos Center Logo header_text
   
 
  
 
 
 
Sponsored by 
ncrr
NIBIB

 
 
MNE

In multimodal imaging, we use both software developed in house and commercial packages. In the Martinos Center, we have developed the MNE software to compute cortically-constrained L2 minimum-norm current estimates and associated dynamic statistical parametric maps from MEG and EEG data, optionally constrained by fMRI. This software includes MEG and EEG preprocessing tools, interactive and batch-mode modules for the forward and inverse calculations, as well as various data conditioning and data conversion utilities.

In addition to the compiled C code tools, MNE Software includes a Matlab toolbox which facilitates access to the fif (functional image file) format data files employed in our software and enables development of custom analysis tools based on the intermediate results computed with the MNE tools. In contrast to other Matlab tools which provide access to fif files, the MNE toolbox is written without resorting to compiled code and is thus composed entirely of m-files.

MNE tools support analysis of data from all three major MEG manufacturers. The Elekta-Neuromag Vectorview and Neuromag-122 data can be directly accessed because we employ the Neuromag native file format in our processing stream. We have also written a conversion utility from the CTF MEG system native format to fif. The 4D Magnes software includes a conversion tool to fif. We have added a utility to incorporate the signals from the compensation data channels as well. The MNE tools are aware of the major noise compensation methods employed for MEG data and apply the necessary corrections in forward calculations and noise estimation procedures as appropriate. Specifically, we support the signal-space projection (SSP) and signal-space separation (SSS) methods employed in Elekta-Neuromag systems and software gradient compensation using reference sensors for CTF and 4D Magnes data.

The MNE tools have been written in C and Matlab. Currently, the software is available for both 32-bit and 64-bit LINUX systems and for Mac OSX (Power PC processors). The LINUX version has been compiled under CentOS and the Mac version under Mac OSX Tiger.