One of the challenges for a multi-site or longitudinal trial is to minimize image variations due to imaging gradient hardware. Imperfections in magnetic field gradient pulses can result in image warping and may limit the detection power of a study attempting to detect structural variations due to pathology. It has been shown that image distortions can significantly affect the quantitative measure of volume, shape, and boundary. The most common source of image distortion is imaging gradient non-linearity, which can distort the images not only in the in-plane directions, but also out of plane.

The Morphometry BIRN is making available a gradient non-linearity distortion correction algorithm, developed at the MGH and UCSD sites, that reduces image distortions and improves test-retest reproducibility of image intensity in multi-site structural MR data studies.

The manuscript from this work has been published:

Jovicich, J, Czanner, S, Greve, D, Haley, E, van der Kouwe, A, Gollub, R, Kennedy, D, Schmitt, F, Brown, G, Macfall, J, Fischl, B, Dale, A. Reliability in multi-site structural MRI studies: Effects of gradient non-linearity correction on phantom and human data. NeuroImage 30:436-443, 2005.

The gradient unwarping software package is available here.

The scripts in this package use spherical harmonics information specific for each scanner gradient set. The tables of spherical harmonics must be obtained by each user from their scanner manufacturer. The scripts do not provide spherical harmonic information in any form (neither the spherical harmonics themselves nor the calculated displacement file) because this is proprietary information. Currently, the user must contact the vendor to get the gradient spherical harmonics information and then use BIRN scripts to create a displacement table (which is used by the image correction software).

The mBIRN makes available to the public a phantom specifically designed to measure the image distortions due to gradient non-linearities. Click here to get information about this phantom.

The gradients supported by this software include:

• Cardiac Resonator Module (CRM) gradient coils from GE Medical Systems (maximum strength = 40 mT/m, slew rate = 150 T/m/s);
• Brain Resonator Module (BRM) gradient coils from GE Medical Systems (22 mT/m, 120 mT/m/ms);
• Sonata gradient coils from Siemens Medical Systems (40 mT/m, 200 T/m/s);
• Avanto gradient coils from Siemens Medical Systems

The application of strong field gradients required for diffusion tensor imaging (DTI) result in eddy currents that induce significant direction-dependent distortions in the resulting images. The Morphometry BIRN is making available a method developed at the Duke site has completed the validation of a method that significantly reduces these distortions in DTI data collected with a standard single-channel coil. This work has been published (Chen B, Guo H, Song AW. Correction for direction-dependent distortions in diffusion tensor imaging using matched magnetic field maps. Neuroimage. 30(1):121-129, 2006). The pulse sequence needed to acquire the data used for the distortion correction is currently implemented on Siemens Trio and GE Horizon platforms and is available through research agreements with the vendors. The GE implementation of the pulse sequence can be distributed to sites that have a research agreement with GE. For the Siemens platform, the pulse sequence will be distributed from their experimental pulse sequence repository using the standard Siemens C2P scheme. Paperwork for the C2P is in progress. Installation and usage literature will be developed and distributed as per usual for Siemens requirements. The executable and source code are available here.

The Large Deformation Diffeomorphic Metric Mapping (LDDMM) algorithm, developed at Johns Hopkins University, aims to quantify metric distances on anatomical structures im medical images. This allows for the direct comparison and quantization of morphometric changes due to, for example, disease or aging.

Aspects of this work have been published:

Faisal Beg, Michael Miller, Alain Trouve, and Laurent Younes. Computing Large Deformation Metric Mappings via Geodesic Flows of Diffeomorphisms. International Journal of Computer Vision, Volume 61, Issue 2; February 2005.

M.I. Miller and A. Trouve and L. Younes, On the Metrics and Euler-Lagrange Equations of Computational Anatomy, Annual Review of biomedical Engineering, 4:375-405, 2002.)

A significant source of error in brain image data is spatial distortions due to inhomogeneities in the main magnetic field (B0), which can arise due to imperfect gradient non-linearity, shimming, magnetic susceptibility effects and chemical shift. These effects can become particularly important for longitudinal studies in which different shim settings can result in substantially different distortions between scan sessions. Although these distortions are most pronounced in functional or diffusion-weighted imaging using echo planar imaging (EPI) sequences, the effect can be significant (on the order of several millimeters) even in conventional structural images. Work by Susumu Mori’s group at the JHU showed that landmark-based distortion correction using the Large Deformation Diffeomorphic Metric Mapping (LDDMM) algorithm can be a viable method to minimize B0 distortion. They have developed software to perform these operations. This software is currently in the testing phase.