Tutorial 2 Converting and preparing data
Introduction
Almost all MRI scanners can output neuroimaging data in DICOM format, a standardized medical image format. Different vendors have developed specific data formats that conform to the DICOM standard and these vendor specific formats are often the default data export type, e.g. PAR/REC on Philips scanners. The analysis of neuroimaging data often starts with converting the vendor-specific data to a data format called NIFTI, since most analysis packages cannot work directly with the original data. The NIFTI format is readable by neuroimaging analysis software such as SPM, FSL or AFNI. Some MRI scanner manufacturers have now implemented the option to export data in NIFTI format directly from the scanner console. This may hopefully make converting data between different formats soon something of the past. If you cannot yet export NIFTI directly from your scanner, there are several free tools around that will convert DICOM and other vendor formats to NIFTI. Below you can find a description and links to several of these tools.
One important thing to realize when using third party tools is that they may change the original data in an unpredictable manner. For example, the orientation of left and right in the image may be flipped in an inappropriate manner, which may even be different for different MRI sequences that were obtained on the same scanner. Make sure to verify the image orientation before starting an experiment or its analysis. This can be easily verified by scanning a phantom with a marker attached or by sticking a vitamin E capsule in the headcoil in a predetermined location.
How to convert data from DICOM to NIFTI
There are several good, free data conversion utilities available on the web that will convert the MRI data from the native scanner format to ANALYZE or NIFTI. Check out the dcm2nii webpage, which also lists several alternatives and the MRIconvert webpage. SPM and Brainvoyager have built-in solutions to convert DICOM data into a format specific for those packages. Often, a site-specific data conversion solution has been programmed.
Tutorial details
This tutorial is still under development. When finished, it will provide a step by step example of the data conversion process based on sample data.
Learning goals
The goal of this tutorial will be twofold. You will learn to convert data to NIFIT format and also how to store data in a structured and safe way.
DICOM to NIFTI
Dcm2nii is a very complete DICOM and Philips PAR/REC converter that will run on Windows, linux and Mac OsX. Dcm2nii is provided as part of an image viewer package called MRIcron. This package is free to download and use and probably your best bet if you do not have an inhouse solution to convert medical images to NIFTI format.
FreeSurfer | mri_convert
Mri_convert is a free general purpose utility for converting data between different file formats. It is distrubuted as part of FreeSurfer and is compatible with 24 formats. Amongst these formats are Siemens, GE and a Generic DICOM format.
SPM + Matlab
SPM has a tool that will allow the conversion of DICOM, MINC and ECAT data format to NIFTI; the default file format in SPM 8. Note that not all types of DICOM are supported. SPM requires Matlab to run.
NITRC

The The Neuroimaging Informatics Tools and Resources Clearinghouse (NITRC) facilitates finding and comparing structural and functional neuroimaging tools and resources. If none of the converters on this page work for you, chances are that you will find one that does work at NITRC
MRI convert
MRI convert will convert data from SIEMENS, PHILIPS and GE scanners to various forms of NIFTI and ANALYZE format. The application is freely available and runs on Windows, Mac OsX and Linux. If DICOM to NIFTI does not work for you, chances are high that MRI convert will.
Matlab 2009A

MATLAB(r) is a high-level language and interactive environment that lets you focus on your course work and applications, rather than on programming details. It enables you to solve many numerical problems in a fraction of the time it takes to write a program in a lower-level language. It is required by SPM.
