Navigation and service

Acceleration of Diffusion MRI in the k-t-Domain with UNFOLD

There have been numerous attempts to reduce MR scanning time without significant degradation of the image quality or loss of information. A properly accelerated sequence can be used to provide a comparable image quality with reduced scan time, a better image quality with the same scan time or combinations of both. This is especially necessary in diffusion MRI where the scan time is intrinsically large and often even too large for clinical applications. Therefore, the sequence acceleration can help make diffusion MRI an even more valuable tool in the diagnosis of a wide range of diseases.
One approach for acceleration that exploits the sparsity of MRI signals in the k-t-domain is Unaliasing by Fourier-Encoding the Overlaps Using the Temporal Dimension (UNFOLD) by Madore et al. [1]. In this method k-space lines are skipped in a way that, for instance, only every second or third line is acquired, leading to an acceleration factor of two or three, respectively. This undersampling of k-space introduces the well-known effect of aliasing in the images (see image in the centre). However, by skipping lines in a carefully chosen pattern, one is able to remove the deliberately introduced artefact during image reconstruction (see image on the right). The resultant accelerated image is of comparable quality but was acquired faster than the standard image (left).
Our group is currently working on UNFOLD acceleration in high angular resolution and q-space diffusion MRI. A successful implementation of UNFOLD in diffusion MRI would open a great prospect for further improving the image quality per scan time in this fascinating technique.

Acceleration of Diffusion MRI in the k-t-Domain with UNFOLDStandard image (left), the accelerated image with aliasing artefacts (centre), the accelerated image with removed artefacts (right).


Bruno Madore, Gary H Glover, and Norbert J Pelc. “Unaliasing by Fourier-Encoding the Overlaps Using the Temporal Dimension (UNFOLD), Applied to Cardiac Imaging and fMRI”. In: Magnetic Resonance in Medicine 42.5 (1999), pp. 813–828.

Additional Information

Contact Person

Prof. N. J. Shah, PhD

Richard Buschbeck
Dr.Seong Dae Yun
Dr. Ana-Maria Oros-Peusquens