You don't have to use the CTF software to convert the DICOM images and to assign the coordinate system. Please do refer to the pictures below for the definition of the fiducial locations, which have not changed.
% read the DICOM files mri = ft_read_mri('single_file_of_DICOM_series.IMA'); % or use a graphical file selection [f, p] = uigetfile('*'); mri = ft_read_mri(fullfile(p, f)); % Making sure you know which side is the right side (e.g. using the vitamin E marker), % assign the nasion (pressing "n"), left ("l") and right ("r") with the crosshairs on % the ear markers. Then finish with "q". cfg = ; cfg.method = 'interactive'; cfg.coordsys = 'ctf'; mri_realigned = ft_volumerealign(cfg,mri); % done!
Following coregistration, you can use the ft_volumereslice function to reslice the MRI, i.e. to interpolate the anatomy onto a new 3D grid that is aligned with the axes of the coordinate system. This prevents problems such as described here.
This manual is written for version 5.4.0 of the CTF-software utilities.
The conversion involves two steps, both using the CTF-program MRIViewer: First, the images are converted from DICOM into the native CTF format. Subsequently, the head coordinate has to be specified, the headshape extracted, a single-sphere model can be fitted to the headshape and multisphere can be constructed (for SAM).
The earphone of the MRI scanner has a large (~1 cm) vitamine E marker built in on the right side. Make sure to check that this marker is visible on the right side.
To keep your data files organized, it is advised that you use a logical directory structure, e.g.:
|/home/…/$subjectcode/||CTF files (.hdm .mri .shape .shapejnfo)|
|analyze/||Analyze files (.hdr .img)|
|dicom/||208 MRI data files (.ima or .dcm)|
|misc/||MRI localizer files (.ima or .dcm)|
Where '$subjectcode' is the coded name of your subject data, for instance: subject_01.
Do the following before starting.the actual conversion procedure:
Now that you have prepared the directory structure for the subject data, you can start with the conversion procedure!
While still in MRIViewer, perform the following actions:
At the DCCN we use ear-molds that come in a variety of sizes to position the MEG head localiser coils just outside the left and right ear . For the MRI we use the same ear-molds, but now with small MRI markers.
On top of that, the right ear also contains a large vitamine E marker to help distinguish left and right. At the location of the nasion we don't put a marker. Below you see three slices with the right-ear marker. Note that the ear marker is the small dot in the middle of the ear shell (yellow arrow), not the large dot close to the ear lobe (red arrow; that is the vitamine E capsule to indicate the right side).
The most elegant way to identify the markers would include:
Transform the head shape somewhat by applying the following things:
Go to file and save the head shape file in the same folder as the .mri is in. The positions should be saved in head coordinates. Saving the headshape to file will create a .shape file and a .shape_info file. Close the brain/head shape extraction dialog by clicking File→close. When asked if the current shape points should be deleted, say No.
In the main window of the MRIViewer, click on fit to head shape” It should give an error of around 5% or less. Save the single-sphere headmodel in the same directory as the .mri is in, this will create an .hdm file (File → Save head model as). You should also save the changes in the MRI file (File, Save). Now you can finally close MRlViewer.
Please note that for most FieldTrip analyses you will not be using a single-sphere volume conduction model, but it might be handy to have the single-sphere head model.
Important to realise is that the multi-sphere model depends on the headshape AND on the position of the gradiometers. Therefore each measurement session (or dataset) should have its own multi-sphere model. The first time you run SAM with new data, SAM asks you to find the .shape file only once. After this is done once other SAM calculations using the same file will automatically get the needed data.