University of Heidelberg
Medical Faculty Mannheim
University Hospital Mannheim
Keine Gesundheitsgefährdung für Frauen mit implantierten kupferhaltigen Verhütungsmitteln bei der MRT,
read more
read more
Job offers now collected in new page on website

Diffusion-weighted imaging of the spinal cord

Michaela Ruttorf and Frank G. Zöllner


By means of magnetic resonance imaging (MRI), descending inhibitory and facilitatory pathways in the spinal cord can be assessed structurally by diffusion tensor imaging (DTI). Because imaging of the spinal cord is challenging [1] due to small cross-sectional dimensions, sensitivity to physiological motion and bone surrounding the spinal cord, such studies are so far rarely reported [2]. We developed a DTI phantom of the lumbar spinal cord to optimise DTI-MRI protocols for use in patients with lower back pain and/or paraplegia.

Materials and Methods

To simulate nerve fibres, a polyamide filament (50 dtex) of 50 μm thickness was selected (Filamentgarn TYPE 611, Trevira GmbH, Bobingen, Germany). The filament was spooled to form a fibre bundle to be placed within a plastic tube. As an emulation of the surrounding spinal segments, a commercially available anatomical model of L1 - L5 (EZ Erler Zimmer, Lauf, Germany) was used. The different fibre phantoms can be easily adjusted and (re)placed into the spinal canal (see Figure 1).

Figure 1: Image of the spinal cord fibre phantom incorporated in an anatomical model of a lumbar spinal cord.

DTI data acquisition was performed on a 3T MAGNETOM Trio whole-body MR scanner (Siemens Medical Solutions, Erlangen, Germany) using a single-shot spin-echo echo planar imaging sequence with following parameters: TR = 5000 ms, TE = 100 ms, FoV = 210 x 20 mm2, matrix size = 178 x 128, voxel size = 2 x 2 x 2 mm3, no gap, bandwidth = 798 Hz/px. Diffusion weighting was performed in multi-directional diffusion weighting (MDDW) mode along 6 non-collinear directions with b = 1000 s/mm2 and repeated 10 times to enhance signal-to-noise ratio. Additionally, a single non-diffusion weighted volume (b = 0 s/mm2) was acquired at each average.


The fibre bundle was cut partially at one side to simulate a spinal cord lesion. The cut is clearly visible at a T1 image of the spinal cord fibre phantom (see Figure 2, left). It is not so clearly visible on a the fractional anisotropy map (see Figure 2, right).

Figure 2: MR images of the spinal cord fibre phantom: T1 image (left) and fractional anisotropy map (right). The area of the lesion is indicated by a white arrow.


[1] Vedantam et al., Neurosurgery, 2014, 74, pp. 1 - 8.
[2] Wang et al., NeuroReport, 2014, 25, pp. 1368 - 1392.

Further reading

  1. J. Rausch, M. Ruttorf, L.R. Schad and F.G. Zoellner. Implementation of a Diffusion Tensor Imaging Phantom of the Lumbar Spinal Cord. Proceedings of the 32nd ESMRMB, 2015, S465 - S467.
  2. J. Andoh, J. Trojan, R. Bekrater-Bodmann, M. Diers, S. Kamping, M. Rance, M. Ruttorf and H. Flor. Structural interhemispheric interactions in primary somatosensory cortices in upper-limb amputees. Organisation for Human Brain Mapping (OHBM) 2014, P-3942.
Contact: Dr. Frank Zöllner last modified: 14.05.2018
to top of page