References:

1

• Introduction
• Diffusion tensor imaging (DTI) has been applied
  to a variety of neurological diseases and brain
  injuries including mild traumatic brain injuries
  (MTBI). One of the problems that this technology
  still faces originates from the fact that it is
  not clear which one of the various quantities
  that can be derived from the diffusion tensors is
  most effective in a given application or
  diagnosis. Among the scalar quantities in use are
  fractional anisotropy (FA) and the mean (MD),
  transversal (Lt) and axial (L1) diffusivity (see
  figure 1), all of which are certain functions of
  the diffusion tensors eigenvalues. Here we
  extend DTI into the temporal domain by comparing
  these measures obtained from subjects who
  suffered a concussion in a sequence of scans with
  control groups who did not suffer from injury.
  In these comparisons we find changes in the
  structural integrity of the white matter during
  the recovery period.

• Results I
• Figure 3 shows the number of significantly
  different voxels in axial slices between the
  concussed subjects and the first control group
  for the first (red), second (green) and third
  (blue) scan. Comparison between the two control
  groups is shown in black.
• For all four quantities there are two distinct
  peaks in the concussed-control comparisons, no
  such peaks exist in control-control
• Around Z115 (cyan region) recovery is seen in
  MD, Lt and L1 but not in FA
• Around Z96 (purple region) recovery is most
  pronounced in FA and L1
• No recovery is seen for FA around Z115 and Lt
  around Z96.

• Results III
• Figure 5 shows regions where significantly
  different voxels are found in the first scan only
  (red), in the first two scans (yellow), in the
  first and third scan (purple) and in all three
  scans (white).
• For all three scans the different voxels are
  located in the same regions
• At Z115 only MD shows changes in greater regions
  bilaterally
• In most cases areas shrink from red to yellow to
  white, indicating recovery. There are few
  locations where the voxels are different only in
  the first and third scan but not the second
  there are no voxels where there is a difference
  in the third scan but not in the first.

• Results II
• Figure 4 shows scaled values for the four
  quantities in voxels that are different for the
  three concussed subjects and the averages for the
  controls.
• FA is significantly increased after concussion in
  all three scans
• MD and Lt are decreased
• Except for the first scan there is no change in
  L1.

• Conclusions
• Recovery in the white matter is found in
  sequences of DTI scans taken within 24h of a
  concussion and about one and two weeks later
• Different quantities derived from the diffusion
  tensors have different sensitivity and
  dependence on location
• The most widely used fractional anisotropy (FA)
  does not seem to be the best diagnostic measure
  as it shows differences between the control
  groups in lower brain areas and no changes in
  time in the upper cortical regions
• Further studies of other quantities, including
  directional information from the tensors
  eigenvectors, are needed to establish DTI
  diagnostic technique for detecting MTBI and
  monitoring recovery from MTBI.

References 1 S.M. Smith et al. (2004) Advances
in functional and structural MR image analysis
and implementation in FSL. Neuroimage
23208-219 2 S.M. Smith et al. (2006)
Tract-based spatial statistics Voxelwise
analysis of multi-subject diffusion data.
Neuroimage 311487-1505
Acknowledgement Work supported by NINDS grant
48299 (JASK).