1. Introduction & Purpose

This tool is intended for those who perform functional MRI on the Siemens Syngo platform and want to
process their functional MRI experiments with Statistical Parametric Mapping (SPM99).

We provide (as is - no guarantees) a tool to convert Syngo exported DICOM images (.MR..*.IMA) to SPM
analyze format. With the currect version it is possible to convert:

• single DICOM images (every slice in a separate DICOM file)
• mosaic DICOM images (n slices in one DICOM file)

2. Hardware/software requirements

Impatient? Go to download.

The conversion program is supported (and also tested) on two operating systems:

• LINUX
• Windows 2000


Note: It should run also on other UNIX systems, if everything is done as for the LINUX system as well as on other Windows versions.

• matlab5 or matlab6
• spm99 (statistical parametric mapping, available from http://www.fil.ion.ucl.ac.uk/spm/)
• medcon program (to convert DICOM images, available from: http://xmedcon.sourceforge.net)
• and some special things depending on your operating system - please see below


For a LINUX/UNIX version you need to have installed:

• mmv (multiple move command)


For Windows

• the DOS version of the UNIX command 'strings' - strings.exe (you may take a Unix 95 Collection Version 7 from http://virtunix.itribe.net/mystuff.html, the package includes also gzip.exe and tar.exe to unzip and untar our dowloadable file)

Uncompress the dowloaded file:

• using DOS/UNIX/LINUX commands:
 
gzip -d dcm2spm.tar.gz
tar -xvf dcm2spm.tar
 
• on Windows PCs you may use the WinZIP program.


The DCM2SPM directory is created in your current directory.

3. Usage

The use of the conversion is rather easy than difficult - via the graphical user interface.

Before running it you need to put your DICOM images on your PC (we do export the data from SIEMENS SONATA and the images are automatically saved as DICOM images).

Organise your DICOM images as follows:

DICOM_dir  -> directory containing all functional MRI data - DICOM images;
 |
 |- MPR    -> sub-directory containing anatomical DICOM images

Launch matlab, start spm99 and behind the matlab prompt type: sonata. The window comes up:

From the pop-up menu select the type of images to be converted. So far, three types of DICOM images are possible convert:

• Single DICOM images
• Mosaic DICOM images (the latest format available from SIEMENS, see description below)

Let's start with some history first, with the previous mosaic format (in fact, the first mosaic format released for the SYNGO platform). Suppose that you run first localizer (three images), i.e. series #1 and then your functional MRI experiment (suppose 125 timepoints, 32 slices), i.e. series #2; you export all the images and this results in the following file naming:

.MR..1.1.year.month.day.a.b.c.x.y.IMA
.MR..1.2.year.month.day.a.b.c.x.y.IMA
.MR..1.3.year.month.day.a.b.c.x.y.IMA
.MR..2.31.year.month.day.a.b.c.x.y.IMA
.MR..3.31.year.month.day.a.b.c.x.y.IMA
.
.
.MR..126.31.year.month.day.a.b.c.x.y.IMA

where a, b, c are most likely time parameters and x, y are the random numbers. The localizer images have all the images named as .MR..1.*.IMA but the functional MRI images have every 32 slices grouped in one mosaic image and the blue number is increasing. The red number is the number of slices decreased by 1. This format has been replaced by the following format and most likely will not be used further (therefore, the conversion routine for this format is not supplied).

The latest mosaic format has some little improvement but it is still not ideal. Again assume that you run first localizer (three images), i.e. series #1 and then your functional MRI experiment (suppose 125 timepoints, 32 slices), i.e. series #2; you export all the images and this results in the following file naming:

.MR..1.1.year.month.day.a.b.c.x.y.IMA
.MR..1.2.year.month.day.a.b.c.x.y.IMA
.MR..1.3.year.month.day.a.b.c.x.y.IMA
.MR..2.31.year.month.day.a.b.c.x.y.IMA
.MR..2.63.year.month.day.a.b.c.x.y.IMA
.MR..2.95.year.month.day.a.b.c.x.y.IMA
.
.
.MR..2.3999.year.month.day.a.b.c.x.y.IMA

where a, b, c are most likely time parameters and x, y are the random numbers. Now the blue number stays at the value of 2 (which is quite logical since we are still in the series #2) but the red number is increasing with the number of slices (=32). However, it would be more convenient to start the red numbers with one (first mosaic image) and increase it by 1 up to the number of timepoints as it is in case of single images (one slice per image). Hopefully, we can have it some day...

Fill in the required fields according to the instructions given by the program and and press the Run conversion button.
After a while you will find your converted images in the SPM directory:

SPM_dir
 |- anat   -> sub-directory with converted anatomical images
 |- ts01   -> time series #1 (default file names im_???.img, can be changed to ts??_???.img)
 |- ts02   -> time series #2
 |- ...    -> etc.

4. Download

If you are interested, you may download here.

Should you have any problems/comments/suggestions please do not hesitate to contact us.