1. health and fitness
2. disease

ORIGINAL CONTRIBUTION
Contribution of White Matter Lesions
to Gray Matter Atrophy in Multiple Sclerosis
Evidence From Voxel-Based Analysis of T1 Lesions in the Visual Pathway
Jorge Sepulcre, MD; Joaquı´n Gon˜i, BS; Joseph C. Masdeu, MD; Bartolome Bejarano, MD;
Nieves Ve´lez de Mendiza´bal, BS; Juan B. Toledo, MD; Pablo Villoslada, MD
Background: The biological basis of gray matter (GM)
atrophy in multiple sclerosis is not well understood, but
GM damage seems to be the most critical factor leading
to permanent disability.
Objective: To assess to what extent white matter (WM)
lesions contribute to regional GM atrophy in multiple
sclerosis.
Design: Because optic pathway GM atrophy and optic
radiation lesions, rather than being related to each
other, could be independent results of the disease, we
applied a nonaprioristic WM method to analyze the
interrelationships of both phenomena. On a voxel-byvoxel basis, we correlated T1 magnetic resonance
imaging–derived lesion probability maps of the entire
brain with atrophy of the lateral geniculate nuclei and
calcarine/pericalcarine cortices.
Setting: Multiple sclerosis center, University of Navarra,
Pamplona, Spain.
Patients: Sixty-one patients with multiple sclerosis.
Main Outcome Measure: Mapping of WM regions contributing to GM atrophy in the optic pathway.
Results: Patients with multiple sclerosis had lateral geniculate nucleus atrophy, which correlated with the presence of lesions specifically in the optic radiations but not
in the rest of the brain. Optic pathway lesions explained
up to 28% of the change of variance in lateral geniculate
nucleus atrophy. Patients also had occipital cortex atrophy, which did not correlate with lesions in the optic radiations or any other WM region.
Conclusions: Focal WM damage is associated with up-
stream GM atrophy, suggesting that retrograde damage
of the perikarya from axonal injury in multiple sclerosis
plaques is one of the significant factors in the genesis of
GM atrophy, although other neurodegenerative processes are probably at work as well.
Arch Neurol. 2009;66(2):173-179
M
ULTIPLE SCLEROSIS
(MS) is a chronic multifocal inflammatory
and degenerative disease of the central nervous system.1 Although it affects predominantly the white matter (WM), it is now
known that the gray matter (GM) is also
involved,2-8 even at early stages.9-12 Different mechanisms underlying GM atrophy
in MS have been postulated. Some of them
Author Affiliations:
Department of Neurology and
Neurosurgery, Clı´nica
Universitaria de Navarra
(Drs Sepulcre, Masdeu,
Bajarano, Toledo, and
Villoslada, Mr Goñi, and
Ms Vélez de Menizábal),
Department of Physics and
Applied Mathematics
(Mr Gon˜i), and Neuroimaging
Laboratory, Center for Applied
Medical Research (Dr Masdeu),
University of Navarra,
Pamplona, Spain.
For editorial comment
see page 159
involve primary damage of the GM, such
as GM plaques that induce demyelination, neuronal loss, and pruning of
dendrites and synapses, or even longdistance effects such as transsynaptic degeneration in areas of normal-appearing
GM.1,13,14 However, GM atrophy could also
be secondary to axonal transection in the
WM plaques, causing wallerian degenera-
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
173
tion with attendant transsynaptic degeneration or a dying-back process affecting
the neuronal perikarya, giving rise to the
transected axon. Many experimental studies have shown that neuronal degeneration often occurs when its axon is transected.15 That WM lesions may induce GM
atrophy is supported by a positive correlation between WM lesion load and whole
GM atrophy.11,16-18 However, it could be
countered that the same mechanisms that
cause greater WM damage may also cause
greater GM damage, independent of axonal transection in the WM. Thus, the real
contribution of WM lesions to regional GM
atrophy remains unclear.19
This issue may be addressed by determining whether there is a specific correlation between volume loss in a given GM
structure and the presence of lesions in the
WM pathways originating from or ending in the same GM structure. In this sense,
the visual pathway could be an excellent
model for studying axonal degeneration
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014
Lateral geniculate nucleus
Optic radiations
Occipital cortex
Figure 1. Experimental approach. The optic pathway is well described anatomically, including the gray matter regions (lateral geniculate nucleus and occipital
cortex) and white matter tracts (pregeniculate pathway and optic radiations) (top). Lesions in multiple sclerosis (in red) are distributed all through the brain,
including the white matter of the optic pathway. By correlating the volume of the lateral geniculate nucleus and occipital cortex with all voxels containing multiple
sclerosis lesions, only lesions transecting axons that originate in or target such gray matter areas will be identified, as is the case with the optic radiations in our
study. Either anterograde or retrograde axonal degeneration can contribute to gray matter atrophy (bottom).
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
174
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014
in MS.20 We hypothesized that lesions of the optic radiations, but not anywhere else in the brain, would correlate with atrophy in the lateral geniculate nucleus (LGN)
and visual cortex in the calcarine sulcus of the occipital
lobe. To test this hypothesis, a method should be used
that avoids an aprioristic knowledge about the relationship among these anatomic structures and queries the entire brain, asking the following question: lesions of what
regions (if any) of the WM of the entire brain correlate
with LGN or calcarine cortex atrophy? To answer this
question while avoiding WM aprioristic biases, we used
lesion probability maps (LPMs)21,22 of the entire brain,
obtained from high-resolution structural magnetic resonance (MR) images and an optimized voxel-based morphometry method8,23 (Figure 1).
Table 1. Characteristics of the Healthy Control
and MS Groups
Age, mean (SD), y
Sex, No. M:F
Duration of MS, mean (SD), y
Disease subtype,
No. CIS/RR/SP/PP
EDSS, median (range)
MSFC, mean (SD)
History of optic neuritis, No.
Signal intensity, mean (SD) b
L lateral geniculate nucleus
R lateral geniculate nucleus
L calcarine cortex
R calcarine cortex
L superior occipital cortex
R superior occipital cortex
L middle occipital cortex
R middle occipital cortex
L inferior occipital cortex
R inferior occipital cortex
METHODS
SUBJECTS
We analyzed a sample of 61 patients who fulfilled the revised
McDonald criteria for the diagnosis of MS.24 They were originally recruited to assess disease activity and retinal changes in
MS.25 All subjects were recruited at the MS Center of the University of Navarra after approval by the ethical review committee of this institution and were studied after having provided
informed consent. Because our study was designed to identify
the relationship between visible WM damage and GM volume
independent of the subtype of the disease, we included patients with all evolution subtypes of MS. Because of the reduced number of patients with the progressive subtype of MS,
we did not compare disease subtypes. The use of immunomodulatory drugs was allowed. Patients were excluded if they
were taking corticosteroids or had had a clinical relapse within
the previous 3 months. A control sample consisted of 20 sexand age-matched healthy controls. Individuals were excluded
from both groups if they had ophthalmologic diseases that might
bias the study (eg, diabetes mellitus or glaucoma). Subject demographics are listed in Table 1.
MR IMAGE ACQUISITION
Magnetic resonance imaging was performed with a 1.5-T imager (Symphony; Siemens, Erlangen, Germany). A 3-dimensional (3D) T1-weighted MR image of the whole head (2-mm
axial sections; repetition time, 2140 milliseconds; echo time,
5.04 milliseconds; 256⫻256 matrix size; field of view, 25 cm;
flip angle, 30°; in-plane resolution, 1⫻1 mm) was acquired for
all subjects. We used T1- rather than T2-weighted images because they more accurately identify the existence of axonal damage in patients with MS.26 In addition, a 3D T1-weighted MR
image with the same acquisition settings and a single dose of
gadolinium (0.1 mmol/kg) (Magnevist; Bayer Schering Pharma
AG, Berlin, Germany) was obtained to identify acute lesions
in the study sample.
MR IMAGING DATA ANALYSIS
Voxel-Based Morphometry Protocol
to Obtain GM Segmented Images
We used a modified version of the optimized voxel-based morphometry protocol23 to obtain the normalized and segmented
GM images from each subject’s T1 MR image, while avoiding
the bias introduced by WM lesions in the normalization and
Control
(n = 20)
MS
(n =61)
37.4 (8.7)
8:12
NA
NA
36.4 (9.0) a
22:39 a
5.7 (7.0)
22/28/5/6
NA
NA
NA
2 (0-7)
0.64 (0.76)
21 (13 R/5 L /
3 bilateral)
0.175 (0.016)
0.161 (0.019)
0.459 (0.056)
0.453 (0.064)
0.310 (0.037)
0.286 (0.030)
0.320 (0.042)
0.300 (0.036)
0.390 (0.055)
0.365 (0.057)
0.141 (0.017) c
0.132 (0.018) c
0.404 (0.052) c
0.395 (0.047) c
0.259 (0.035) c
0.224 (0.029) c
0.271 (0.031) c
0.233 (0.030) c
0.317 (0.043) c
0.249 (0.036) c
Abbreviations: CIS, clinically isolated syndrome; EDSS, Expanded
Disability Status Scale; L, left; MS, multiple sclerosis; MSFC, MS Functional
Composite; NA, not applicable; R, right; RR, relapsing-remitting; PP, primary
progressive; SP, secondary progressive.
a No significant differences between groups.
b Signal intensity values (adimensional eigenvalues) extracted from each
gray matter region of interest.
c P ⬍ .001.
segmentation procedures.8,27 For this purpose we used statistical parametric mapping (SPM2) software (Wellcome Department of Cognitive Neurology, University College of London,
London, England; http://www.fil.ion.ucl.ac.uk/spm) running
under MATLAB version 6.5 (MathWorks Inc, Natick, Massachusetts). To preserve the total within-voxel volume, all voxel
signal intensities in the final GM segmented images were multiplied by the Jacobian determinants ( Jacobian modulation) derived from the spatial normalization.23 Finally, images were
smoothed with a 12-mm full-width at half-maximum isotropic gaussian filter to increase the signal-to-noise ratio and to
account for variations in normal gyral anatomy.23
White Matter LPMs
To obtain the WM LPMs, 2 independent observers outlined at
voxel level all of the WM lesions on each of the transverse sections of the 3D T1-weighted MR image of each patient by means
of MRIcro software (Chris Rorden, PhD, University of Nottingham, Nottingham, England; http://www.sph.sc.edu/comd
/rorden/mricro.html). Interrater reliability was excellent (intraclass correlation coefficient=0.89; P⬍.001).25 In this first
step, we obtained a 3D binary mask for each patient. We then
normalized each lesion mask from a native space to a stereotactic space by using the SPM2 software and the settings obtained from the previous voxel-based morphometry procedure (see the preceding section). Finally, to approximate the
neuropathological characteristics of MS with greater tissue damage in the center of the lesions, lesion masks were smoothed
and converted to LPMs by applying an isotropic gaussian kernel (12-mm full-width at half-maximum).22 Therefore, we obtained 3D LPMs where each voxel had a probability value of
being classified as lesion (ranging from 0 to 1), with a higher
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
175
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014
A
C
+8
+ 10
+ 12
+ 14
+ 16
+ 18
+ 20
+ 22
+ 24
+ 26
+ 28
+ 30
+ 32
+ 34
+ 36
+ 44
0.6
0.4
0.2
0.0
+ 38
+ 40
+ 42
D
E
B
+8
+ 10
+ 12
+ 14
+ 16
+ 18
+ 20
+ 22
+ 24
+ 26
F
G
+ 28
+ 30
+ 32
+ 34
+ 36
+ 38
+ 40
+ 42
+ 44
+ 46
Figure 2. Lesion probability map for a single patient, shown in color (A) and in gray scale overlapped with a T1 magnetic resonance imaging template (B). Also
shown are gray matter regions of interest used in the analysis of the visual system: lateral geniculate nucleus (C), calcarine cortex (D), inferior occipital cortex (E),
middle occipital cortex (F), and superior occipital cortex (G).
value at the center of the lesion than in the periphery (Figure 2A
and B).
STATISTICAL ANALYSIS
As we focused our study on the visual system, a set of brain
structures with well-defined anatomy, we used a region of interest (ROI)–level analysis for the assessment of GM atrophy
by means of the Marsbar automated parcellation method.28 Specifically, we studied the LGNs, calcarine cortices (Brodmann
area 17/18), and pericalcarine cortices (Brodmann area 18/19)
of the occipital lobe, divided into superior, inferior, and middle
occipital cortex ROIs (Figure 2C-G). All normalized ROIs were
obtained from Marsbar (occipital cortex ROI) and Voitool (LGN
ROI) (http://www.ihb.spb.ru/~pet_lab/VTO/VTOMain.html)
SPM toolboxes. The GM volumes in patients with MS and in
healthy controls were compared by means of the SPSS 13.0 statistical package (SPSS Inc, Chicago, Illinois). The normal distribution of the variables was assessed with the KolmogorovSmirnov test. To search for correlations between GM atrophy
and WM lesions, we used the framework of the general linear
model implemented in the SPM2 software. We applied voxelby-voxel multiple linear regressions to correlate the WM LPMs
(of the entire brain; nonaprioristic WM analysis) and the GM
ROI extracted data. This analysis strategy allowed us to look
into the WM of the entire brain for lesions associated with GM
atrophy, while avoiding the introduction of any a priori knowledge of possible WM anatomic target locations (such as the optic radiations). All correlations were adjusted by (1) sex and
age and (2) the regional predilection of MS lesions, which, for
instance, tend to cluster in the periventricular regions. The last
adjustment was made by using a WM regional lesion frequency mask, created with the average of the individual 3D binary masks. The use of this mask ensured that only voxels of
WM were included in the analysis, excluding from the analysis any GM voxels highlighted by the smoothing protocol used
to create the LPM. The mean signal intensity values (adimensional eigenvalues) of each significant cluster were extracted
with the VOI (volume of interest) toolbox of SPM2 to show
the global R, R2, and P values of every correlation. The level of
significance for the results was set at P⬍.05 after correction
for multiple comparisons to minimize type I error (false discovery rate method).29 To achieve consistent results, only clusters with 100 or more voxels were retained.30
RESULTS
To evaluate the contribution of WM lesions to GM atrophy in MS, we first calculated the GM volume reduction in the optic pathway (LGN and occipital cortex) in
patients compared with controls. After that, locations of
WM lesions associated with GM reduction were identified by a voxel-based lesional approach. This method allowed us to evaluate the correlation between MS lesion
location and GM reduction (Figure 1) while avoiding any
a priori knowledge about anatomic WM target locations.22,31 Finally, we calculated the percentage of the variance in the GM atrophy explained by the presence of pre-
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
176
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014
A
20
F
10
0
L
R
L
R
B
F
10
5
0
Figure 3. Axial projections showing white matter lesion locations correlated with lateral geniculate nuclei volumes (color bar indicates F statistic values). A, Right
lateral geniculate nucleus volume decrease was specifically correlated with white matter lesions in the right optic radiations. B, Left lateral geniculate nucleus
volume decrease was specifically correlated with white matter lesions in the left optic radiations, although this cluster was not retained after multiple-test
correction.
viously identified WM lesions. Compared with healthy
controls, patients with MS had smaller right and left LGN
(P⬍.001 in both cases; Table 1).
In the WM lesions analysis, we found that LGN volume loss correlated with the presence of lesions in the
optic radiations in both the right (Figure 3A) and left
(Figure 3B) hemispheres. No other WM lesions, either
within or outside of the optic pathway, correlated with
LGN atrophy. The presence of WM lesions in the right
optic radiations explained 28% of variance changes for
right LGN atrophy, and the presence of WM lesions in
the left optic radiations explained 13% of the variance
change for left LGN atrophy (Table 2). Despite the anatomic specificity of the results in the left hemisphere, this
cluster did not remain significant after multiplecomparison correction. No significant difference was
found in LGN volume between patients with and without a history of optic neuritis. Twelve subjects had gadolinium-enhancing lesions, but none of them involved the
optic radiations.
Finally, we also found that patients with MS had significant volume decrease in both right and left occipital
cortices (P ⬍ .001 in all cortical ROIs: calcarine, superior occipital, inferior occipital, middle occipital; Table 1).
However, we did not find any correlation between occipital cortex volume and the presence of WM plaques
anywhere in the brain.
COMMENT
Different mechanisms have been proposed to explain GM
damage in MS.1,13,14 Transsynaptic degeneration has received particular attention in the visual system. Lesions
in the optic nerve are associated with transsynaptic
changes in the LGN,32,33 particularly in the parvocellular layer,32 and in the visual cortex.34 Ciccarelli et al34 have
shown reduced connectivity in the optic radiations, measured by tractography-based mapping, 1 year after an epi-
Table 2. Correlation Between Lateral Geniculate Nucleus
Volume and White Matter Lesions
White Matter Lesion Location a
Right Optic
Radiations
No. of voxels
P value
Uncorrected
With FDR correction
F statistic value
Maximum coordinates x, y, z
Rb
R 2 (%) b
2815
⬍.001
.02
26.23
36, −36, 15
−0.529
0.280 (28)
Left Optic
Radiations
1977
.002
.06
10.04
−27, −66, 9
−0.368
0.136 (13)
Abbreviation: FDR, false discovery rate.
a Locations are correlated with right and left lateral geniculate nucleus,
respectively.
b Derived from the mean signal intensity values (eigenvalues) of each
significant cluster.
sode of optic neuritis, suggesting not only transsynaptic
effects but also axonal loss in the optic radiations corresponding to neuronal loss in the LGN. However, the effect
of optic radiation lesions on the volume of the LGN and
occipital cortex has not been previously studied, to our
knowledge.
Our results indicate that the presence of WM lesions
in the optic radiations, originating in the LGN, explains
up to 28% of the change of variance for the LGN total
volume, reinforcing the concept that retrograde damage
of the neurons (perikarya) due to injury of their axons
in the MS plaques contributes significantly to GM atrophy. It is important that the WM of the entire brain was
tested against LGN atrophy, by means of a WM nonaprioristic approach. Only lesions in either of the optic
radiations, and not anywhere else in the brain, were associated with LGN atrophy. However, the lesions explained more volume variance change on the right side
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
177
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014
than the left side probably because our combined sample
had a greater lesion load in the right hemisphere.
In summary, because our method prevented any bias
introduced by the prior knowledge of anatomic correlations among WM structures, the finding of an association between LGN atrophy and lesions in the optic
radiations of both hemispheres is robust and could confirm our hypothesis that lesions in the optic radiations
cause atrophy in the LGN. However, because more than
70% of LGN volume variance remains unexplained, the
effect of optic radiation lesions on the development of
GM atrophy is limited and other factors clearly could
play a role in the genesis of LGN atrophy. It is possible
that lesions of the WM not visible on MR images
(normal-appearing WM changes), pregeniculate
lesions32-34 (although we did not find any differences in
LGN volume between patients with and without previous optic neuritis), or GM in situ processes may contribute to the rest of the variance.
Finally, patients with MS had highly significant atrophy of the occipital cortex that was not explained by the
presence of WM lesions in its related pathways. This result is in concordance with previously published data
using magnetic transfer ratio in the visual cortex of patients with and without lesions in the optic radiations.34
To explain this apparent paradox, it could be argued that
the occipital cortex is a highly connected area and therefore the impact of axonal damage in the optic radiations
was not enough to cause atrophy. Loss in connectivity,
however, may not be the only cause of the occipital cortex atrophy observed in MS. The convergence of the results from different MR imaging techniques, showing the
lack of correlation with lesions in the optic radiations,
suggests the existence of other neurodegenerative or GM
in situ processes.
In conclusion, our results suggest that neuronal loss
from retrograde degeneration due to axonal damage within
MS plaques could be one of the significant factors in the
genesis of GM atrophy. Because this mechanism explains less than half of the variance, other neurodegenerative processes are probably at work as well.
Accepted for Publication: May 4, 2008.
Correspondence: Pablo Villoslada, MD, MS Center, Departamento de Neurologı´a, Clı´nica Universitaria de Navarra, Universidad de Navarra, Pı´o XII 36, 31008 Pamplona, Navarra, Spain (pvilloslada@unav.es).
Author Contributions: All authors had full access to all
of the data in the study and take responsibility for the
integrity of the data and the accuracy of the data analysis. Study concept and design: Sepulcre, Masdeu, and
Villoslada. Acquisition of data: Villoslada. Analysis and interpretation of data: Sepulcre, Gon˜i, Masdeu, Bejarano,
Ve´lez de Mendiza´bal, Toledo, and Villoslada. Drafting of
the manuscript: Sepulcre, Masdeu, Toledo, and Villoslada.
Critical revision of the manuscript for important intellectual content: Gon˜i, Masdeu, Bejarano, Ve´lez de Mendiza´bal,
and Villoslada. Statistical analysis: Sepulcre, Gon˜i, and
Toledo. Obtained funding: Villoslada. Administrative, technical, and material support: Toledo. Study supervision: Gon˜i,
Masdeu, Bejarano, Ve´lez de Mendiza´bal, and Villoslada.
Financial Disclosure: None reported.
Funding/Support: This study was supported by the European Union (grant 512146LSH-2003-1.2.2.-2 to Dr Masdeu), the Spanish Ministry of Health (grant CM#05/
00222 to Dr Sepulcre; FIS#PI052520 to Dr Masdeu; and
FIS#PI051201 to Dr Villoslada), the Navarra government (Mr Gon˜i and Dr Masdeu), the Basque Country government (Ms Ve´lez de Mendiza´bal), the foundation “UTE
project CIMA” (Dr Masdeu), and the Fundacio´n Uriach
(Dr Villoslada).
Additional Contributions: We thank the MS Society of
Navarra and the subjects who kindly agreed to take part
in this study.
REFERENCES
1. Hauser SL, Oksenberg JR. The neurobiology of multiple sclerosis: genes, inflammation, and neurodegeneration. Neuron. 2006;52(1):61-76.
2. Filippi M. Multiple sclerosis: a white matter disease with associated gray matter
damage. J Neurol Sci. 2001;185(1):3-4.
3. Sailer M, Fischl B, Salat D, et al. Focal thinning of the cerebral cortex in multiple
sclerosis. Brain. 2003;126(pt 8):1734-1744.
4. Dalton CM, Chard DT, Davies GR, et al. Early development of multiple sclerosis
is associated with progressive grey matter atrophy in patients presenting with
clinically isolated syndromes. Brain. May 2004;127(pt 5):1101-1107.
5. Rovaris M, Bozzali M, Iannucci G, et al. Assessment of normal-appearing white
and gray matter in patients with primary progressive multiple sclerosis: a diffusiontensor magnetic resonance imaging study. Arch Neurol. 2002;59(9):14061412.
6. Dehmeshki J, Chard DT, Leary SM, et al. The normal appearing grey matter in
primary progressive multiple sclerosis: a magnetisation transfer imaging study.
J Neurol. 2003;250(1):67-74.
7. Davies GR, Altmann DR, Hadjiprocopis A, et al. Increasing normal-appearing grey
and white matter magnetisation transfer ratio abnormality in early relapsingremitting multiple sclerosis. J Neurol. 2005;252(9):1037-1044.
8. Sepulcre J, Sastre-Garriga J, Cercignani M, Ingle GT, Miller DH, Thompson AJ.
Regional gray matter atrophy in early primary progressive multiple sclerosis: a
voxel-based morphometry study. Arch Neurol. 2006;63(8):1175-1180.
9. Chard DT, Griffin CM, Parker GJ, Kapoor R, Thompson AJ, Miller DH. Brain atrophy in clinically early relapsing-remitting multiple sclerosis. Brain. 2002;
125(pt 2):327-337.
10. Sastre-Garriga J, Ingle GT, Chard DT, Ramio´-Torrentà L, Miller DH, Thompson
AJ. Grey and white matter atrophy in early clinical stages of primary progressive
multiple sclerosis. Neuroimage. 2004;22(1):353-359.
11. De Stefano N, Matthews PM, Filippi M, et al. Evidence of early cortical atrophy in
MS: relevance to white matter changes and disability. Neurology. 2003;60(7):
1157-1162.
12. Bjartmar C, Trapp BD. Axonal and neuronal degeneration in multiple sclerosis:
mechanisms and functional consequences. Curr Opin Neurol. 2001;14(3):271278.
13. Brück W. Inflammatory demyelination is not central to the pathogenesis of multiple sclerosis. J Neurol. 2005;252(suppl 5):v10-v15.
14. Perry VH, Anthony DC. Axon damage and repair in multiple sclerosis. Philos Trans
R Soc Lond B Biol Sci. 1999;354(1390):1641-1647.
15. Quarantelli M, Ciarmiello A, Morra VB, et al. Brain tissue volume changes in relapsing-remitting multiple sclerosis: correlation with lesion load. Neuroimage.
2003;18(2):360-366.
16. Prinster A, Quarantelli M, Orefice G, et al. Grey matter loss in relapsingremitting multiple sclerosis: a voxel-based morphometry study. Neuroimage. 2006;
29(3):859-867.
17. Charil A, Dagher A, Lerch JP, Zijdenbos AP, Worsley KJ, Evans AC. Focal cortical atrophy in multiple sclerosis: relation to lesion load and disability. Neuroimage.
2007;34(2):509-517.
18. Pirko I, Lucchinetti CF, Sriram S, Bakshi R. Gray matter involvement in multiple
sclerosis. Neurology. 2007;68(9):634-642.
19. Frohman EM, Costello F, Stüve O, et al. Modeling axonal degeneration within the
anterior visual system: implications for demonstrating neuroprotection in multiple sclerosis. Arch Neurol. 2008;65(1):26-35.
20. Lee MA, Smith S, Palace J, et al. Spatial mapping of T2 and gadolinium-
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
178
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014
21.
22.
23.
24.
25.
26.
27.
enhancing T1 lesion volumes in multiple sclerosis: evidence for distinct mechanisms of lesion genesis? Brain. 1999;122(pt 7):1261-1270.
Charil A, Zijdenbos AP, Taylor J, et al. Statistical mapping analysis of lesion location and neurological disability in multiple sclerosis: application to 452 patient data sets. Neuroimage. 2003;19(3):532-544.
Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, Frackowiak RS.
A voxel-based morphometric study of ageing in 465 normal adult human brains.
Neuroimage. 2001;14(1, pt 1):21-36.
Polman CH, Reingold SC, Edan G, et al. Diagnostic criteria for multiple sclerosis:
2005 revisions to the “McDonald Criteria.” Ann Neurol. 2005;58(6):840-846.
Sepulcre J, Murie-Fernandez M, Salinas-Alaman A, Garcı´a-Layana A, Bejarano
B, Villoslada P. Diagnostic accuracy of retinal abnormalities in predicting disease activity in MS. Neurology. 2007;68(18):1488-1494.
Comi G, Rovaris M, Leocani L, Martinelli V, Filippi M. Assessment of the damage of the cerebral hemispheres in MS using neuroimaging techniques. J Neurol Sci. 2000;172(suppl 1):S63-S66.
Esteban FJ, Sepulcre J, de Mendiza´bal NV, et al. Fractal dimension and white
matter changes in multiple sclerosis. Neuroimage. 2007;36(3):543-549.
Tzourio-Mazoyer N, Landeau B, Papathanassiou D, et al. Automated anatomical
labeling of activations in SPM using a macroscopic anatomical parcellation of
the MNI MRI single-subject brain. Neuroimage. 2002;15(1):273-289.
28. Genovese CR, Lazar NA, Nichols T. Thresholding of statistical maps in functional neuroimaging using the false discovery rate. Neuroimage. 2002;15(4):
870-878.
29. Grossman M, McMillan C, Moore P, et al. What’s in a name: voxel-based morphometric analyses of MRI and naming difficulty in Alzheimer’s disease, frontotemporal dementia and corticobasal degeneration. Brain. 2004;127(pt 3):
628-649.
30. Bates E, Wilson SM, Saygin AP, et al. Voxel-based lesion-symptom mapping.
Nat Neurosci. 2003;6(5):448-450.
31. Evangelou N, Konz D, Esiri MM, Smith S, Palace J, Matthews PM. Size-selective
neuronal changes in the anterior optic pathways suggest a differential susceptibility to injury in multiple sclerosis. Brain. 2001;124(pt 9):1813-1820.
32. Goldby F. A note on transneuronal atrophy in the human lateral geniculate body.
J Neurol Neurosurg Psychiatry. 1957;20(3):202-207.
33. Audoin B, Fernando KT, Swanton JK, Thompson AJ, Plant GT, Miller DH. Selective magnetization transfer ratio decrease in the visual cortex following optic neuritis.
Brain. 2006;129(pt 4):1031-1039.
34. Ciccarelli O, Toosy AT, Hickman SJ, et al. Optic radiation changes after optic neuritis detected by tractography-based group mapping. Hum Brain Mapp. 2005;
25(3):308-316.
(REPRINTED) ARCH NEUROL / VOL 66 (NO. 2), FEB 2009
179
WWW.ARCHNEUROL.COM
©2009 American Medical Association. All rights reserved.
Downloaded From: http://archpedi.jamanetwork.com/ on 10/15/2014