Brain Viscoelasticity Alteration in Chronic-Progressive
Kaspar-Josche Streitberger1, Ingolf Sack1*, Dagmar Krefting1, Caspar Pfu ¨ller2, Ju ¨rgen Braun3,
Friedemann Paul2,5., Jens Wuerfel2,4.
1Department of Radiology, Charite ´ – University Medicine Berlin, Berlin, Germany, 2NeuroCure Clinical Research Center and Experimental and Clinical Research Center,
Charite ´ – University Medicine Berlin and Max Delbrueck Center for Molecular Medicine, Berlin, Germany, 3Institute of Medical Informatics, Charite ´ – University Medicine
Berlin, Berlin, Germany, 4Institute of Neuroradiology, University Luebeck, Luebeck, Germany, 5Clinical and Experimental Multiple Sclerosis Research Center, Charite ´ -
University Medicine Berlin, Berlin, Germany
Introduction: Viscoelastic properties indicate structural alterations in biological tissues at multiple scales with high
sensitivity. Magnetic Resonance Elastography (MRE) is a novel technique that directly visualizes and quantitatively measures
biomechanical tissue properties in vivo. MRE recently revealed that early relapsing-remitting multiple sclerosis (MS) is
associated with a global decrease of the cerebral mechanical integrity. This study addresses MRE and MR volumetry in
chronic-progressive disease courses of MS.
Methods: We determined viscoelastic parameters of the brain parenchyma in 23 MS patients with primary or secondary
chronic progressive disease course in comparison to 38 age- and gender-matched healthy individuals by multifrequency
MRE, and correlated the results with clinical data, T2 lesion load and brain volume. Two viscoelastic parameters, the shear
elasticity m and the powerlaw exponent a, were deduced according to the springpot model and compared to literature
values of relapsing-remitting MS.
Results: In chronic-progressive MS patients, m and a were reduced by 20.5% and 6.1%, respectively, compared to healthy
controls. MR volumetry yielded a weaker correlation: Total brain volume loss in MS patients was in the range of 7.5% and
1.7% considering the brain parenchymal fraction. All findings were significant (P,0.001).
Conclusions: Chronic-progressive MS disease courses show a pronounced reduction of the cerebral shear elasticity
compared to early relapsing-remitting disease. The powerlaw exponent a decreased only in the chronic-progressive stage of
MS, suggesting an alteration in the geometry of the cerebral mechanical network due to chronic neuroinflammation.
Citation: Streitberger K-J, Sack I, Krefting D, Pfu ¨ller C, Braun J, et al. (2012) Brain Viscoelasticity Alteration in Chronic-Progressive Multiple Sclerosis. PLoS ONE 7(1):
Editor: Wang Zhan, University of Maryland, United States of America
Received October 17, 2011; Accepted December 8, 2011; Published January 20, 2012
Copyright: ? 2012 Streitberger et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the German Research Foundation (Exc 257 to F.P. and C.P. and Sa 901/3 to I.S.). The funders had no role in study design,
data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
. These authors contributed equally to this work.
Magnetic resonance imaging (MRI) has emerged as most
important paraclinical tool for the diagnosis and monitoring of
disease activity in multiple sclerosis (MS), which is reflected by the
current MS diagnostic criteria [1,2]. However, disease specificity
of conventional MRI parameters, such as T2 lesion load, is limited
 and their association with clinical course and neurological
disability is only modest . Consequently, considerable scientific
effort is necessary to further improve the diagnostic specificity and
predictive value of MRI in MS.
In a recent development of medical imaging, the sensitivity of
viscoelastic constants to the hierarchy of the mechanical matrix of
tissue is exploited. Mechanical imaging, also called elastography
, promises sensitivity to pathological processes which affect the
tissue architecture across a continuum of microscopical to
macroscopical scales. To date, only MR elastography (MRE) 
is capable of measuring the viscoelastic properties of living brain in
its natural environment [7,8,9]. In cerebral MRE, harmonic
vibrations are applied to the skull; the induced shear waves inside
the brain are subsequently captured by motion sensitive MRI
sequences. Thus, the obtained frequency-resolved complex wave
images can be analyzed by wave inversion for viscoelastic
parameter recovery. Multifrequency MRE combined with two-
parameter springpot analyses have recently been applied in brain
studies of aging [10,11], normal pressure hydrocephalus [12,13]
and MS . The reported springpot parameters m and a describe
the viscoelastic dispersion of the complex shear modulus G* of
brain tissue by a powerlaw, i.e. a linear function in logarithmical
coordinates. The observed linear increase of log(G*) over log(v),-
with v being the angular drive frequency, implies scaling of G*
and therewith a hierarchical order of the mechanical network in
PLoS ONE | www.plosone.org1 January 2012 | Volume 7 | Issue 1 | e29888
brain tissue . The shear elasticity m is related to the inherent
strength or integrity of this network, while the powerlaw exponent
a is related to geometry, i.e. the topology or fractal dimension of
the network [15,16]. Both springpot parameters are potentially
sensitive markers of a variety of neurological disorders affecting the
global mechanical scaffold of brain at multiple scales.
Our primary hypothesis is based on a recently published study
of multifrequency MRE applied to a group of patients suffering
from relapsing-remitting MS . At this early state of MS, a
significant decrease of m compared to healthy controls was
observed while a remained unchanged. In a preliminary
interpretation of these results, it was speculated that the network
integrity of the brain is affected due to widespread neurodegen-
erative processes in MS, whereas the geometry of this network is
not. We therefore hypothesize that chronic inflammation in
progressive MS causes a further progression of brain matrix
degradation, resulting in a reduction of both, m and a. Our
secondary hypothesis is that the parameters m and a may reflect
neurodegenerative processes in MS due to the sensitivity of
biomechanical parameters to structural changes also on the
microscopic level. Thus, MRE parameters may yield a higher
sensitivity to neurodegenerative processes in comparison to, e.g.
MR volumetry .
To test both hypotheses, patients with chronic-progressive MS
disease course and matched healthy volunteers were enrolled in a
cross-sectional study, combining multifrequency MRE and MR
volumetry. Supplementary data were acquired by standard clinical
and neuroradiological methods. We compared our resulting data
to literature values in relapsing-remitting forms of MS. Thus,
conclusions could be drawn about the impact of early relapsing-
remitting and chronic-progressive MS disease courses on brain
Twenty-three consecutive patients with primary (n=6), or
secondary (n=17) chronic progressive disease course were
recruited for this study from the NeuroCure Clinical Research
Center or the neuroimmunology outpatient clinic of the
Experimental and Clinical Research Center. Demographic data
are summarized in Table 1. The study was approved by the ethics
committee of the Charite ´ – University Medicine Berlin, and all
participants gave written informed consent. A cohort of thirty-
eight age and gender matched healthy volunteers without
neurological or psychiatric diseases served as controls.
Magnetic resonance elastography
MRI measurements were performed on a 1.5 tesla scanner
(Sonata, Siemens Medical Systems, Erlangen, Germany). A
detailed scheme of the experiment is presented in Figure 1. The
MRE protocol comprised a single-shot spin-echo echo planar
imaging sequence with a sinusoidal motion-encoding gradient
(MEG) in through-plane direction (z-direction) that was used to
acquire three transversal image slices in a central slab through the
cerebrum (number of MEG cycles: 4, MEG amplitude: 35 mT/m,
repetition time [TR]: 3.0 s, echo time [TE]: 149 ms, field of view
[FoV]: 1926192 mm2, matrix size: 1286128, slice thickness:
For each slice, the acquisition was repeated 64 times during one
period of 80 ms. A toggle gradient with alternating sign of motion
sensitization was applied with a stepwise increasing delay between
vibration onset and motion-encoding of 2.5 ms. A multifrequency
vibration with maximum amplitude of approximately 0.5 mm in
parallel direction to the long axis of the magnet was fed into an
actuator by a carbon fiber piston. The resulting time-resolved wave
images, u(x,y,t) (with x and y as spatial coordinates) were Fourier-
transformed for decomposition into complex wave images at drive
frequency: U(x,y,v); v=2pf with f being the drive frequency.
U(x,y,v) was subjected to spatial filtering to suppress compression
wave components and noise using isotropic lower (upper) thresholds
of 5.56 (50.0) m21, 8.33 (66.67) m21, 10.0 (90.9) m21, and 10.0
(100.0) m21for f=25 Hz, 37.5 Hz, 50 Hz and 62.5 Hz, respec-
tively. These bandpass filters were applied to the masked wave
images excluding air outside the brain but including CSF and
ventricles. Complex modulus images were obtained by wave
inversion (G*(x,y,v)=2rv2U/DU), with D as the 2D-Laplace
operator and r being the tissue’s density of 1 kg/dm3), spatially
averaged within the segmented brain parenchyma. The resulting
global modulus function was fitted by a least-square routine. A good
match between model and multifrequency data was reached by a
combination of Voigt and Maxwell models given by the three-
parameter Zener model. However, thelattermodel incorporated an
additional parameter – a second shear modulus – rendering the
interpretation of viscoelastic constants rather cumbersome. The
optimal trade-off between physical significance and representation
of the frequency dependency of our data was achieved by a two-
parameter springpot model G*=k(iv)a, that represents a powerlaw
and interpolates between springs and dashpots introducing a
fractional element k=m(12a)ga. Two springpot variables are
representedbym and a,whileg isthe viscosityofthemedium,which
has to be defined a priori in order to derive shear moduli comparable
Table 1. Demographical data, clinical characteristics, brain
volumes, brain parenchymal fraction (BPF) and viscoelastic
constants m and a according to the springpot model.
groupMS (sp)/MS (pp) controlMS (rr)*control*
N 17 (sp)/6 (pp) 38 4534
female N9 (sp)/4 (pp) 222317
age52 (9.1)/51 (5.0) 48 (9.7)38 (8.0)37 (11.4)
Mean EDSS 5.6 (1.3)/5.3 (1.8)0 1.6 (1.4)0
V in dm3
1.513 (0.178)1.636 (0.068)n.a.n.a.
m (in kPa)2.607 (0.482)3.278 (0.314)3.025 (0.459)3.545 (0.556)
0.2756 (0.0108) 0.2934 (0.0086) 0.2937 (0.0129) 0.2928 (0.0131)
MS – multiple sclerosis; sp – secondary progressive; pp – primary progressive; rr
– relapsing remitting; EDSS – expanded disability status scale; n.a. – not
applicable; standard deviations are given in brackets;
*data taken from  and processed corresponding to the data of progressive
Brain Viscoelasticity in Progressive MS
PLoS ONE | www.plosone.org2January 2012 | Volume 7 | Issue 1 | e29888
to other brain mechanical tests. In  it was suggested to use
g=3.7 Pa?s for human brain, which is the viscosity value of the
brainparenchymaderived bytheZenermodel,and was alsoused in
this study. Every participant underwent MRE of approximately
15 minutes duration additionally to a routineclinical MRIprotocol:
T2- and proton density-weighted images (TR: 5780 ms, TE1:
13 ms, TE2: 81 ms, TE3: 121 ms; 3 mm slice thickness, no gap, 44
contiguous axial slices). Bulk white matter lesion load of T2-
weighted scans were routinely measured using the MedX v.3.4.3
software package, as described previously .
Assessment of brain volume and calculation of the brain
Total brain tissue volumes, normalised for subject head size,
were estimated on three-dimensional T1-weighted sequences (TR:
2110 ms, TE: 4.38 ms, inversion time [TI] 1100 ms, flip angle
15u, resolution 1 mm3), applying SIENAX [19,20,21], part of FSL
. SIENAX extracted brain and skull images from single whole-
head input data . The brain image was then affine-registered
to MNI152 space [23,24], using the skull image to determine the
registration scaling. This was primarily in order to obtain the
volumetric scaling factor to be used as normalization for head size.
Subsequently, tissue-type segmentation with partial volume
estimation was carried out  to calculate the total volume of
the brain tissue, including separate estimates of volumes of grey
matter, white matter, peripheral gray matter and ventricular CSF.
Brain parenchymal fraction (BPF) is expressed as ratio of the sum
of white and grey brain matter by intracranial volume.
All data were analyzed using IBM SPSS Statistics 18 (IBM
Corporation, Route 100, Somers, NY, USA). Correlation between
clinical parameters, MRI and MRE was assessed by Spearman’s
correlation coefficient. Differences between patients and controls
regarding age, MRI and MRE parameters were analyzed by the
Mann-Whitney U test. Multivariate linear regression analyses
Figure 1. Scheme of cerebral multifrequency MRE. a: The MRI scanner is combined with a device for acoustical head stimulations comprising:
1) a signal generator that produces a multifrequency signal composed from four harmonic frequencies of 25, 37.5, 50 and 62.5 Hz; 2) a loudspeaker
for generating acoustic vibrations; 3) an extended piston that transfers the vibrations into the scanner and 4) a head cradle for stimulating head
vibrations mainly along the head-feet direction. b: A single-shot echo planar imaging (EPI) sequence is sensitized to harmonic motions by a 60-Hz
sinusoidal motion encoding gradient (MEG) of four cycles and directed through-plane. The image planes are positioned in transverse orientation
through the brain (parallel to the ‘‘anterior and posterior commissure line (AC-PC)’’) in a central slab of the brain. The resulting wave images display
the motion component along the head-feet direction corresponding to the major vibration direction of the actuator. c: Image processing comprises
Fourier decomposition of the superposed oscillations yielding four complex single-frequency wave images, corresponding to the experimentally
applied vibration frequencies. Each of the wave images is separately inverted, resulting in four complex-valued shear modulus images, whose values
are averaged within a region of interest comprising the parenchyma within the image slice (demarcated in the wave images by white lines).
Brain Viscoelasticity in Progressive MS
PLoS ONE | www.plosone.org3 January 2012 | Volume 7 | Issue 1 | e29888
were performed to assess the influence of MRI and MRE and
clinical parameters on the springpot parameter m. Owing to the
study design, all tests should be understood as constituting
exploratory data analysis, such that no adjustments for multiple
testing were made. Receiver operating curve (ROC) analyses were
performed in Matlab (MATLAB 6.5 R13, The MathWorks,
Natick, MA, USA). Preliminary cut-off values for m and a were
attained by the maximum Youden index.
Reduced brain volumes in MS patients compared to
In accordance to the literature , MS patients exhibited
significantly lower brain parenchymal volumes expressed as
compared to healthy controls, corresponding to an average
reduction of 7.5% (P,0.001) in patients.
BPF in MS patients (0.9610) was 1.69% lower than in healthy
controls (0.9775; P,0.001). As these relative values depend on the
grey matter/white matter contrast, we were not able to reliably
compare our present data to a previous study on relapsing-
remitting MS patients that was performed on a different MR
The volume- and BPF data are plotted in figure 2. Group values
are given in table 1.
Reduction of viscoelasticity constants in MS patients
Both springpot constants m and a were significantly lower in MS
patients compared to healthy controls, corresponding to average
reductions of 20.46% and 6.07%, respectively (P always,0.001).
The group response is plotted in figure 3. Averaged values of m and
a as well as their disease related alterations are tabulated.
Correlation of viscoelasticity constants with age and
In the MS group, the shear elasticity m did neither correlate with
age (R=0.133, P=0.545), nor with the BPF (R=0.220,
P=0.313). Also no correlation was seen between m and the total
brain volume (R=0.375, P=0.077). Also a was not correlated to
age (R=20.210, P=0.335), BPF (R=0.017, P=0.936) and
volume (R=0.021, P=0.924). In this group, no correlation was
seen between age and brain volume (R=0.323, P=0.133) and age
and BPF (R=20.118, P=0.593). In the group of healthy
volunteers, m correlated with age (R=20.607, P,0.001) and
volume (R=0.481, P=0.002) while between m and BPF only a
trend was discernable (R=0.309, P=0.059). In agreement to
MRE studies in the literature, a does not change with age in
healthy volunteers (R=20.124, P=0.459) [10,11], neither it does
with volume(R=0.104, P=0.535)
P=0.482). Within the patient group, there was no correlation
between number or volume of whole brain T2 lesions with any of
the viscoelasticity parameters (data not shown). However, five
datasets needed to be removed from the semi-automated T2
analysis due to movement artefacts, resulting in a small number of
datasets, warranting a repetition in a larger trial.
In a multivariate linear regression analysis including age, BPF
and total brain volume as covariates, none of these variables
independently predicted elasticity parameter m in the MS group
(standardized coefficients Beta=0.044, P=0.845; Beta=0.214,
P=0.320; Beta=0.355, P=0.122, respectively).
Combining both groups, m is correlated with age (R=20.337,
P=0.008), brain volume(R=0.559,
(R=0.502, P,0.001). a displays only a trend with age
(R=20.249, P=0.053) but is correlated to brain volume
(R=0.333, P=0.009) and BPF (R=0.394, P=0.002).
P,0.001) and BPF
Correlation of MRE and MRI parameters with clinical data
Not surprisingly, EDSS (expanded disability status scale)
correlated with disease duration (R=0.506, P=0.014), but not
with BPF and total brain volume (R=0.02, P=0.928; R=0.05,
P=0.822, respectively). BPF inversely correlated with disease
duration (R=20.418, P=0.047).
m was neither correlated with EDSS (R=20.077, P=0.728)
nor with disease duration (R=20.190, P=0.384). When
including EDSS and disease duration as additional covariates in
Figure 2. Brain atrophy in MS patients. Significantly reduced brain parenchymal volume (a) and brain parenchymal fraction (BPF) (b) in MS
patients compared to matched healthy individuals (*** P,0.001). The boxplots depict the lower and upper quartiles as well as the 50thpercentile
(median). Full data range is presented by the whiskers. sp – secondary progressive, pp – primary progressive, rr – relapsing remitting.
Brain Viscoelasticity in Progressive MS
PLoS ONE | www.plosone.org4 January 2012 | Volume 7 | Issue 1 | e29888
the multivariate linear regression analysis, none of these were
independent predictors of m. Similarly, a was neither correlated
with EDSS (R=0.099, P=0.653) nor with disease duration
MRE differentiates healthy from diseased brains
In an area under the receiver operating characteristics curve
analysis (AUROC), we found a high value for both the shear
elasticity m (0.896) and the powerlaw exponent a (0.936) for the
discrimination of MS patients versus controls (Fig. 4). Preliminary
cut-off values for diagnostic MRE in MS are proposed with
m=2.677 kPa and a=0.285. The corresponding sensitivity and
specificity values are 0.896 and 0.658 and 0.936 and 0.842 for m
and a, respectively.
We investigated brain viscoelastic properties in patients with
progressive disease courses of MS, and compared these to a group
of healthy individuals of matched age and gender. The most
important findings of our work are that (i) both, the shear elasticity
m and the powerlaw exponent a differentiate MS from healthy
controls and (ii), although there is an association between brain
atrophy and changes in viscoelasticity in the total study cohort,
brain volume reduction is not the only contributor to the
alterations of viscolelastic properties in MS. It has recently been
demonstrated in a large group of healthy volunteers that brain
atrophy only weakly contributes to viscoelastic constants measured
by MRE . The correlation between viscoelastic constants and
atrophy found in our study is not surprising, since parenchymal
degradation - physiologically during aging or pathologically
accelerated in MS - impacts both, the mechanical matrix of brain
and its volume. However, correlation does not necessarily imply
causality. The fact that viscoelastic constants measured by MRE
can represent volume independent markers is indicated by a
higher rate of parameter changes in MRE compared to MR
volumetry. However, at this state of research we cannot definitely
exclude interactions between volumetrical changes, brain geom-
etry and MRE. As a consequence, we have been focusing on
averaged viscoelastic constants related to the whole parenchyma
visible in the slab of the chosen image slices. As such, cerebral
MRE still represents a global method providing data for
widespread effects occult to standard imaging modalities. In MS,
the measured springpot constants suggest that large regions of the
central cerebrum (i.e. the region of our wave image slices) are
Figure 3. Reduction of brain parenchymal viscoelastic constants. MS patients present with significantly reduced brain parenchymal elasticity
m (a, P,0.001), but also with a reduction in the powerlaw exponent a (b, P,0.001) in MS patients with progressive disease course. The boxplot
depicts the lower and upper quartiles as well as the 50thpercentile (median). Full data range is presented by the whiskers. sp – secondary progressive,
pp – primary progressive, rr – relapsing remitting; *data for rr-MS are taken from  and reprocessed according to the methods reported in herein.
Figure 4. Viscoelastic constants for the detection of brain
pathology. Individual data of shear elasticity m and powerlaw
exponent a of brain tissue in healthy volunteers and MS patients. The
areas under the receiver characteristics curve (AUROC) for separating
healthy volunteers from MS patients are 0.896 and 0.936 for m and a,
Brain Viscoelasticity in Progressive MS
PLoS ONE | www.plosone.org5 January 2012 | Volume 7 | Issue 1 | e29888
mechanically degraded due to neuroinflammation. The reduction
of the shear elasticity reported in  for early MS is enhanced in
chronic-progressive MS (212.7% versus 220.5%, see table 1).
Most interestingly, a - which was not significantly correlated in
early relapsing-remitting MS - is altered in chronic-progressive
disease courses (0.29% versus 6.1%). This draws our attention to
the structure sensitivity of m and a. The principal relationship
between powerlaw exponent a and the hierarchical architecture of
biomechanical networks was illustrated in  by numerical
simulations and multifrequency MRE experiments of skeletal
muscle. It was stated there that an increase in the fractal dimension
of the network (given in contracting muscle by the establishment of
myosin cross bridges) yields to an increase of the powerlaw
exponent a while m is influenced by the network-inherent spring
constants quantifying the mechanical integrity of the underlying
tissue. Translating these findings to brain viscoelasticity in MS
leads us to the following hypothesis: In early disease stages the
integrity of the mechanical matrix is degraded while its inherent
geometrical order is preserved. However, this order becomes
affected during further disease progression, resulting in a
continuous decrease of m, but furthermore also to a reduction in
a. The latter process may occur either in a slow and continuous
manner, or by discontinuous remodelling of the tissue similar to a
phase transition from an ordered to a disordered state of the brain
parenchyma. To date it is not entirely clear which kind of
mechanical cerebral tissue structure determines m or a. Recent in
vivo MRE experiments in mouse models suggested that both
demyelination and inflammation contribute to the observed
deterioration of the cerebral mechanical scaffold [27,28].
Our study is limited by a relatively small number of patients,
preventing an individual group analysis of secondary progressive
and primary progressive MS. A further limitation lies in different
protocols used between previous studies published in [10,14,29],
and our current set-up [11,12,13]. We therefore reprocessed the
data of  using the filter limits and the masking of wave images
as given in the Methods section. Furthermore, the alignment of
image slices in a peripheral slab of the brain through the upper
part or slightly above the ventricles (as done in [10,14,29]) may
result in a larger variation of viscoelastic constants than observed
in a more central position as an effect of enhanced wave scattering.
Our conclusive statements regarding early relapsing-remitting and
chronic-progressive MS are therefore limited to relative effects
between healthy controls and patients groups. Applying three-
dimensional multifrequency MRE in future studies will alleviate
current restrictions to the slice alignment and may therewith help
to establish generalized viscoelastic thresholds for diagnostic
applications of MRE.
In summary, we demonstrate that cerebral viscoelastic constants
are reduced in chronic-progressive MS. In contrast to early
relapsing-remitting MS disease courses, the springpot powerlaw
exponent a is reduced in chronic-progressive MS, indicating a loss
of the mechanical network geometry due to chronic neuroin-
flammation. MRI volumetry is less sensitive to changes in chronic-
progressive MS and represents therewith only a minor predictor
for viscoelastic constants measured by MRE.
We thank our study nurses Cordula Rudolph, Franziska Lipske, and Antje
Conceived and designed the experiments: K-JS IS FP JW. Performed the
experiments: K-JS IS. Analyzed the data: K-JS IS DK CP JB FP JW.
Contributed reagents/materials/analysis tools: IS JW DK JB. Wrote the
paper: K-JS IS FP JW.
1. McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, et al. (2001)
Recommended diagnostic criteria for multiple sclerosis: guidelines from the
International Panel on the diagnosis of multiple sclerosis. Ann Neurol 50: 121–127.
2. Montalban X, Tintore M, Swanton J, Barkhof F, Fazekas F, et al. (2010) MRI
criteria for MS in patients with clinically isolated syndromes. Neurology 74:
3. Charil A, Yousry TA, Rovaris M, Barkhof F, De Stefano N, et al. (2006) MRI
and the diagnosis of multiple sclerosis: expanding the concept of ‘‘no better
explanation’’. Lancet Neurol 5: 841–852.
4. Barkhof F (2002) The clinico-radiological paradox in multiple sclerosis revisited.
Curr Opin Neurol 15: 239–245.
5. Ophir J, Cespedes I, Ponnekanti H, Yazdi Y, Li X (1991) Elastography: a
quantitative method for imaging the elasticity of biological tissues. Ultrason
Imaging 13: 111–134.
6. Muthupillai R, Ehman RL (1996) Magnetic resonance elastography. Nature
Med 2: 601–603.
7. Sack I, Beierbach B, Hamhaber U, Klatt D, Braun J (2008) Non-invasive
measurement of brain viscoelasticity using magnetic resonance elastography.
NMR Biomed 21: 265–271.
8. Kruse SA, Rose GH, Glaser KJ, Manduca A, Felmlee JP, et al. (2008) Magnetic
resonance elastography of the brain. Neuroimage 39: 231–237.
9. Green MA, Bilston LE, Sinkus R (2008) In vivo brain viscoelastic properties
measured by magnetic resonance elastography. NMR Biomed 21: 755–764.
10. Sack I, Beierbach B, Wuerfel J, Klatt D, Hamhaber U, et al. (2009) The impact
of aging and gender on brain viscoelasticity. Neuroimage 46: 652–657.
11. Sack I, Streitberger KJ, Krefting D, Paul F, Braun J (2011) The influence of
physiological aging and atrophy on brain viscoelastic properties in humans.
PlosOne 6: e23451.
12. Streitberger KJ, Wiener E, Hoffmann J, Freimann FB, Klatt D, et al. (2011) In
vivo viscoelastic properties of the brain in normal pressure hydrocephalus. NMR
Biomed 24: 385–392.
13. Freimann FB, Streitberger KJ, Klatt D, Lin K, McLaughlin J, et al. (2011)
Alteration of brain viscoelasticity after shunt treatment in normal pressure
hydrocephalus. Neuroradiology;2011 May 3. [Epub ahead of print].
14. Wuerfel J, Paul F, Beierbach B, Hamhaber U, Klatt D, et al. (2010) MR-
elastography reveals degradation of tissue integrity in multiple sclerosis.
Neuroimage 49: 2520–2525.
15. Gurtovenko AA, Blumen A (2005) Generalized Gaussian structures: Models for
polymer systems with complex topologies. Polymer Analysis, Polymer Theory.
Berlin: Springer. pp 171–282.
16. Klatt D, Papazoglou S, Braun J, Sack I (2010) Viscoelasticity-based Magne-
tic resonance elastography of skeletal muscle. Phys Med Biol 55: 6445–
17. Miller DH, Barkhof F, Frank JA, Parker GJ, Thompson AJ (2002) Measurement
of atrophy in multiple sclerosis: pathological basis, methodological aspects and
clinical relevance. Brain 125: 1676–1695.
18. Wuerfel J, Bellmann-Strobl J, Brunecker P, Aktas O, McFarland H, et al. (2004)
Changes in cerebral perfusion precede plaque formation in multiple sclerosis: a
longitudinal perfusion MRI study. Brain 127: 111–119.
19. Smith SM (2002) Fast robust automated brain extraction. Hum Brain Mapp 17:
20. Smith SM, De Stefano N, Jenkinson M, Matthews PM (2001) Normalized
accurate measurement of longitudinal brain change. J Comput Assist Tomogr
21. Smith SM, Zhang Y, Jenkinson M, Chen J, Matthews PM, et al. (2002)
Accurate, robust, and automated longitudinal and cross-sectional brain change
analysis. Neuroimage 17: 479–489.
22. Smith SM, Jenkinson M, Woolrich MW, Beckmann CF, Behrens TE, et al.
(2004) Advances in functional and structural MR image analysis and
implementation as FSL. Neuroimage 23 Suppl 1: S208–219.
23. Jenkinson M, Smith S (2001) A global optimisation method for robust affine
registration of brain images. Med Image Anal 5: 143–156.
24. Jenkinson M, Bannister P, Brady M, Smith S (2002) Improved optimization for
the robust and accurate linear registration and motion correction of brain
images. Neuroimage 17: 825–841.
25. Zhang Y, Brady M, Smith S (2001) Segmentation of brain MR images through a
hidden Markov random field model and the expectation-maximization
algorithm. IEEE Trans Med Imaging 20: 45–57.
26. Sastre-Garriga J, Ingle GT, Chard DT, Cercignani M, Ramio-Torrenta L, et al.
(2005) Grey and white matter volume changes in early primary progressive
multiple sclerosis: a longitudinal study. Brain 128: 1454–1460.
27. Schregel K, Wuerfel E, Garteiser P, Prozorovskiy T, Merz H, et al.
Biomechanical properties quantified in vivo by Magnetic Resonance Elasto-
graphy correlate with myelination and brain parenchymal properties - a
Brain Viscoelasticity in Progressive MS
PLoS ONE | www.plosone.org6 January 2012 | Volume 7 | Issue 1 | e29888
combined 7 Tesla MRE and histopathology study in a mouse model of Multiple Download full-text
Sclerosis; 2011 April; Montreal. 40 p.
28. Riek K, Hamann I, Millwald J, Pfueller C, Hirsch S, et al. MR elastography of
mice in experimental autoimmune encephalitis; 2011 April; Montreal, 1476.
29. Klatt D, Hamhaber U, Asbach P, Braun J, Sack I (2007) Noninvasive assessment
of the rheological behavior of human internal organs using multifrequency MR
elastography: A study of brain and liver viscoelasticity. Phys Med Biol 52:
Brain Viscoelasticity in Progressive MS
PLoS ONE | www.plosone.org7 January 2012 | Volume 7 | Issue 1 | e29888