Diffusion Tensor Imaging in Late Posttraumatic Epilepsy
The main objective of this study was to use diffusion tensor imaging (DTI) to search and quantify the extent of abnormality beyond the obvious lesions seen on the T2 and fluid-attenuation inversion recovery (FLAIR) magnetic resonance images in patients with chronic traumatic brain injury (TBI) with and without epilepsy. DTI was performed on 23 chronic TBI patients (with late posttraumatic epilepsy, n=14; without epilepsy, n=9) and 11 age-matched controls. The ratios of fractional anisotropy (FA) and mean diffusivity (MD) between the regions of interest beyond the T2/FLAIR-visualized abnormality and the corresponding contralateral normal-appearing region were calculated. FA and MD ratios were compared for relative changes in these parameters among the TBI subjects with and without epilepsy and controls. Tissue volumes exhibiting abnormalities on DTI also were measured in these patients. The mean regional FA ratio was significantly lower, whereas the mean regional MD value was higher in patients with TBI compared with controls. The mean regional FA ratio was significantly lower in TBI patients with epilepsy (0.57+/-0.059) than in those without epilepsy (0.68+/-0.039). Although the regional MD ratio was higher in TBI patients with epilepsy (1.15+/-0.140) relative to those without epilepsy (1.09+/-0.141), the difference did not reach statistical significance. The tissue volume with low FA value also was found to be higher in TBI patients with epilepsy than without. Severity of injury as predicted by the DTI-derived increased volume of microstructure damage is associated with delayed posttraumatic epilepsy in TBI patients. These findings could be valuable in predicting epileptogenesis in patients with chronic TBI.
Epilepsia, 46(9):1465–1471, 2005
Blackwell Publishing, Inc.
2005 International League Against Epilepsy
Diffusion Tensor Imaging in Late Posttraumatic Epilepsy
Rakesh K. Gupta,
Sona Saksena, †Atul Agarwal, §Khader M. Hasan, ‡Mazhar Husain,
‡Vikas Gupta, and §Ponnada A. Narayana
Department of Radiodiagnosis, Sanjay Gandhi Postgraduate Institute of Medical Sciences, and †Departments of Neurology and
‡Neurosurgery, King George’s Medical University, Lucknow, India; and §Department of Diagnostic and Interventional Imaging,
University of Texas Medical School at Houston, Houston, Texas, U.S.A.
Summary: Purpose: The main objective of this study was to
use diffusion tensor imaging (DTI) to search and quantify the
extent of abnormality beyond the obvious lesions seen on the
and fluid-attenuation inversion recovery (FLAIR) magnetic
resonance images in patients with chronic traumatic brain injury
(TBI) with and without epilepsy.
Methods: DTI was performed on 23 chronic TBI patients (with
late posttraumatic epilepsy, n =14; without epilepsy, n =9) and
11 age-matched controls. The ratios of fractional anisotropy (FA)
and mean diffusivity (MD) between the regions of interest be-
yond the T
/FLAIR-visualized abnormality and the correspond-
ing contralateral normal-appearing region were calculated. FA
and MD ratios were compared for relative changes in these pa-
rameters among the TBI subjects with and without epilepsy and
controls. Tissue volumes exhibiting abnormalities on DTI also
were measured in these patients.
Results: The mean regional FA ratio was significantly lower,
whereas the mean regional MD value was higher in patients
with TBI compared with controls. The mean regional FA ratio
was significantly lower in TBI patients with epilepsy (0.57 ±
0.059) than in those without epilepsy (0.68 ± 0.039). Although
the regional MD ratio was higher in TBI patients with epilepsy
(1.15 ±0.140) relative to those without epilepsy (1.09 ±0.141),
the difference did not reach statistical significance. The tissue
volume with low FA value also was found to be higher in TBI
patients with epilepsy than without.
Conclusions: Severity of injury as predicted by the DTI-
derived increased volume of microstructure damage is asso-
ciated with delayed posttraumatic epilepsy in TBI patients.
These findings could be valuable in predicting epileptogenesis
in patients with chronic TBI. KeyWords: Diffusion tensor
imaging—Traumatic brain injury—Posttraumatic epilepsy—
Fractional anisotropy—Magnetic resonance imaging.
Traumatic brain injury (TBI) is one of the most common
causes of morbidity and mortality in developed countries,
and posttraumatic epilepsy (PTE) is the major sequela
(1,2). PTE is classified into three groups, depending on
when seizures occur after TBI: (a) within 24 h, (b) during
the first week, and (c) after >1 week (3,4). The first two
groups are termed early PTE and are the result of direct re-
sponse to the brain damage. The third category represents
late PTE. The severity and type of brain injury appear to
correlate with the incidence of PTE. The factors that are
known to contribute to the risk of PTE are duration of
unconsciousness, dura penetration, degree of direct cor-
tical damage, and genetic predisposition to epilepsy (5).
Structural damage resulting from trauma itself, or hypoxic
damage and subsequent scarring and inadequate blood
flow also are the major contributing factors to the risk of
PTE. The investigation and management of patients after
Accepted April 19, 2005.
Address correspondence and reprint requests to Dr. R.K. Gupta at
MR Section, Department of Radiodiagnosis, Sanjay Gandhi Post Grad-
uate Institute of Medical Sciences, Lucknow, UP, India, 226014. E-mail:
head injury must include the accurate and complete iden-
tification of cerebral damage (6). In patients with PTE,
magnetization transfer imaging (MTI) has been used for
visualization and characterization of diffuse axonal injury
that appears normal on T
-weighted images (7,8). In the
early phase of trauma, reactive axonal swelling and myelin
sheath collapse and fragmentation occur while reactive
gliosis dominates in the delayed posttraumatic period (9).
It has been shown that the extent of gliosis visualized on
MTI beyond the T
-visible abnormality correlates with
the intractability in PTE (10).
Water diffusion in tissues provides an additional de-
gree of contrast in magnetic resonance imaging (MRI).
This contrast depends on the microscopic motion of water,
and therefore, diffusion MRI has the potential to visualize
pathology at a microscopic level that may not be evident on
conventional MRI, as demonstrated in stroke by Mosley
et al. (11) in their seminal work and more recently in other
neurologic diseases (12–15).
Water molecules exhibit preferential diffusion along
certain directions in tissues because of the presence of
membranes and other structures that restrict the molecular
1466 R. K. GUPTA ET AL.
diffusion. This directional dependence or anisotropy can
be exploited to gain information about the tissue organi-
zation at a microscopic level (16). Determination of the
complete diffusion tensor is realized by acquiring MRI
data with diffusion gradients applied along at least six
non-collinear directions and one data set acquired in the
absence of diffusion gradients (or weak gradients) (17,18).
The two commonly used rotationally invariant scalar pa-
rameters that are derived from diffusion tensor imag-
ing (DTI) are the mean diffusivity (MD) and fractional
anisotropy (FA). MD, the trace of the diffusion matrix,
is the average measure of the molecular diffusion and is
affected by the cellular size and integrity (17,19). FA is a
measure of the diffusion anisotropy and has a value be-
tween 0 (completely isotropic) and 1 (highly anisotropic).
FA reflects the degree of alignment of cellular structures
within the fiber tracts and their structural integrity (20).
Because glial proliferation results in structurally dense
and disorganized tissue, gliosis results in decreased FA
and increased MD. Complete characterization of the dif-
fusion tensor can potentially provide important and spe-
cific pathologic information. Thus DTI offers increased
sensitivity over conventional MRI in assessing the mi-
crostructural damage in brain parenchyma (21,22). DTI
has been shown to be useful in study of diseases, such as
acute stroke (23), cerebral ischemia (24), traumatic brain
injury (25), schizophrenia (26), multiple sclerosis (27),
and epilepsy (28–31).
The aim of this study was to investigate the utility of DTI
measures in demonstrating difference in the microstruc-
tural changes in the normal-appearing tissue beyond the
conventional imaging-visible abnormality in chronic TBI
patients with and without epilepsy and evaluating the pre-
dictive role of quantitative regional FA and MD in epilep-
MATERIALS AND METHODS
Twenty-three patients with documented chronic closed
TBI with or without seizures who were seen in the De-
partment of Neurosurgery and Neurology, KG’s Medical
University, Lucknow, India, were evaluated with conven-
tional MRI and DTI. All these patients had history of
trauma for a period of >1 year at the time of inclusion
in this study. These patients were either first seen after 1
year (n = 14) or were seen on a regular follow-up after
initial head injury. Patients with retained metallic foreign
bodies or scalp plate were excluded from the study. After
a careful review of the patient’s medical history, detailed
physical and neurologic examination [a routine electroen-
cephalogram (EEG)] were performed. The 19 male and
four female patients ranged in age from 7 to 50 years
(mean, 28.5 years). Fourteen patients with epilepsy had
their first episode of seizure between 1 and 3 years (mean,
2.1 years) after TBI. Nine patients with TBI had no history
of seizure on a regular follow-up after head injury, varying
from 1 to 5 years. Of 14 patients with PTE, eight had par-
tial with or without secondarily generalized seizures, and
the other six had generalized tonic–clonic seizures. All pa-
tients with PTE were taking phenytoin sodium (PHT; 200–
400 mg/day, eight as monotherapy), five patients were also
taking clobazam (CLB; 10–25 mg/day), and four patients
were taking valproate (VPA; 600–1,200 mg/day). Serum
levels of antiepileptic drugs (AEDs) were not determined.
Seizures were controlled by AEDs in 11 patients. In the
remaining patients, seizures were considered intractable
(n = 3) when they could not be controlled even after tak-
ing two or three AEDs for ≥6 months. EEG abnormalities
seen in 12 patients included slow-wave transients, spo-
radic spikes or sharp waves, and continuous slow-wave
activity over the damaged brain.
All patients were imaged between 1 to 3 years after
head injury. In addition, normative data were obtained in
11 healthy age-matched controls with no history of TBI
or any other neurologic disorder. Informed consent was
obtained from all the subjects and/or nearest kin.
Whole-brain conventional MRI and DTI were acquired
on a 1.5-Tesla GE MRI scanner (General Electric Medi-
cal Systems, Milwaukee, WI, U.S.A.) by using a standard
quadrature birdcage receiver and transmit radio frequency
head coil. The maximum available gradient amplitude
was33mT/m, and the slew rate of 120 T/m/s. The
conventional MRI protocol included T
spin-echo (FSE) images with repetition time (TR)/echo
time (TE)/echo train length (ETL)/number of excitations
(NEX) of 6,000/200/16/4, and spin-echo (SE) images with
TR/TE/NEX of 1,000/14/2. Both T
-weighted and T
weighted images were acquired from axial sections of 3
mm slice thickness with no interslice gap, 240 mm field-
of-view (FOV), and matrix size of 256 × 256.
DTI was acquired by using a single-shot echo planar
dual spin-echo sequence with ramp sampling (32). The
dual spin-echo–echo sequence reduces image distortions
in the diffusion-weighted images by compensating for the
effect of eddy currents (33). The sequence used a spectral-
selective 90-degree pulse for fat suppression. The number
of diffusion gradient pulses in a dual spin-echo sequence is
four. The durations, δ
,ofthese gradient pulses are δ
=17 ms. The effective diffusion time is
∼50 ms (TE/2). The b-factor was set to 1,000 s/mm
8s;TE, 100 ms; and NEX, 8. In total, 34–38 axial sections
were acquired with a slice thickness of 3 mm, no interslice
gap, FOV of 240 mm. The acquisition matrix was 128 ×
80, and a homodyne algorithm was used to construct to
128 × 128. This was zero filled to reconstruct an image
matrix of 256 ×256. A balanced (34), rotationally invari-
ant (35,36) dodecahedral diffusion-encoding scheme with
10 uniformly distributed directions over the unit sphere
Epilepsia, Vol. 46, No. 9, 2005
DTI IN LATE POSTTRAUMATIC EPILEPSY 1467
was used for generating the DTI data. To enhance signal-
to-noise ratio (SNR) and reduce phase fluctuations, the
magnitude constructed images were repeated (NEX of 8)
and temporally averaged by the scanner software.
DT-MRI data analysis
The magnitude-averaged data were transferred to a
workstation for further analysis. In general, the DTI data
analysis involves three major steps: preprocessing, pro-
cessing, and postprocessing.
The raw images were cropped and stripped by using a
semiautomated procedure to remove the scalp for isolat-
ing the brain. The DWI data were spatially filtered with
a3× 3 median filter. The data were then distortion cor-
rected for shear, scale, rotation, and translation by using
the DT-MRI toolbox (37,38) that incorporated the two-
dimensional perspective models in the Automated Image
and Registration (AIR) package (39).
The distortion-corrected data were interpolated to attain
isotropic voxels and decoded to obtain the tensor field for
each voxel. The tensor field data were then diagonalized by
using the analytic diagonalization method (35) to obtain
the eigenvalues (λ
) and the three orthonormal
). The tensor field data and
eigenvalues were used to compute the mean diffusivity
(Eq. 1) and fractional anisotropy (Eq. 2) for each voxel.
Data postprocessing and quantification
Data processing and analysis were performed by us-
ing an in-house–developed DTI-Toolbox implemented un-
der IDL (Research Systems Inc., Boulder, CO, U.S.A.;
http://rsinc.com/). To facilitate the region-of-interest
(ROI) placement for quantitative analysis, the DTI-derived
maps were displayed and overlaid on images with differ-
ent contrasts (T
-weighted, FLAIR, DWI, ADC, etc.) in
the three orthogonal planes for visual inspection. Ellipti-
cal (Fig. 1), rectangular and free hand ROIs, encompass-
ing the areas within and outside the T
lesions, were placed by an experienced neuroradiologist
(R.K.G.) on the T
and FLAIR images in an unbiased
way for quantifying FA and MD. ROIs also were placed
on the corresponding contralateral, normal-appearing re-
gions. The size of the ROI varied from 5 ×5 pixels to 7 ×7
pixels. Then the FA and MD ratios of the region within and
and FLAIR abnormality and the corresponding
contralateral normal-appearing region were calculated in
these patients. In chronic TBI patients, visual inspection
of the FA maps showed abnormal area beyond the visible
abnormality, and this helped us to restrict to this area only.
In controls, ROIs also were placed in both frontal and oc-
cipital periventricular white matter to quantify the FA and
MD ratio for the purpose of comparison. Volumes were
calculated from the number of pixels within the ROI, and
all the slices showing trauma were included.
A Student’s t test was performed to determine the statis-
tically significant differences between the ROI measure-
ments in chronic TBI patients with and without epilepsy
and in controls. Difference in the FA and MD ratios be-
tween the normal-appearing region beyond T
abnormality and the corresponding contralateral normal-
appearing region were compared with controls. A p value
of <0.05 was considered to be statistically significant.
On MRI, parenchymal injury/trauma was located in the
frontal lobe (n =10), parietal lobe (n =1), occipital lobe (n
= 2), and frontoparietal region (n =1) in chronic TBI pa-
tients with epilepsy (n =14). In patients without epilepsy,
the injury/trauma was present in frontal lobe (n =5), tem-
poral lobe (n =2), occipital lobe (n =1), and frontoparietal
(n = 1). All 11 healthy age-matched controls with no his-
tory of TBI showed normal distribution of white matter
anisotropy (see Fig. 1). In all chronic TBI patients with-
out seizures (Fig. 2) and with seizures (Fig. 3), the FA
map showed the extent of reduced FA beyond the T
The regional mean FA and mean MD ratios in these
patients are summarized in Table 1. The mean FA ratio
in the region beyond T
and FLAIR abnormality was sig-
nificantly lower in all TBI patients compared to controls.
The regional mean FA ratio beyond the abnormality was
significantly lower in chronic TBI patients with epilepsy
compared with patients without epilepsy. The mean FA ra-
tio in the region within the T
and FLAIR abnormality was
significantly lower in patients compared with controls.
However, the regional mean FA ratio did not reach the
level of statistical significance within the T
abnormality in chronic TBI patients with epilepsy com-
pared with patients without epilepsy.
The MD ratio in the region beyond T
and FLAIR ab-
normality was significantly higher in all patients com-
pared with controls. Although the regional MD ratio was
higher in chronic TBI patients with epilepsy compared
with chronic TBI patients without epilepsy, it did not
reach the level of statistical significance beyond the T
and FLAIR abnormality. The mean MD ratio in the region
within the T
and FLAIR abnormality was significantly
Epilepsia, Vol. 46, No. 9, 2005
1468 R. K. GUPTA ET AL.
FIG. 1. A 26-yearold normal control. A: T
. B: Apparent. diffusion coefﬁcient (ADC). C: Fractional anisotropy (FA). D: Color-coded FA
fused with ADC through the normal ventricles shows normal distribution of white matter anisotropy. Note the elliptical regions of interests
placed in both frontal lobes (C) to calculate FA and mean diffusivity (MD) ratios.
higher in patients compared with controls. The volume
in the region beyond the T
and FLAIR abnormality was
higher in chronic TBI patients with epilepsy (range, 7–
11.4 cc) compared with chronic TBI patients without
epilepsy (range, 4.2–7.1 cc).
In the present study, we observed an area of abnor-
mality on FA maps beyond the T
- and FLAIR-visible
pathology. This suggests that the injury extended beyond
the conventional MRI-visible abnormality. The low FA
value is consistent with large microstructural disorganiza-
tion secondary to occult cerebral damage in patients with
chronic TBI without and with epilepsy. This observed low
FA ratio was found to be significantly lower in the patients
with seizure compared with that in the nonseizure group,
suggesting increased gliosis in TBI with epilepsy. It is
known that severity of injury is directly responsible for
the increase in delayed PTE (40). In the present series,
DTI-derived microstructural brain damage was more se-
vere in chronic TBI patients with epilepsy compared with
FIG. 2. Ayoung man age 24 years was on regular follow-up for the last 2 years after a closed head injury. Imaging was performed at 2
years after injury to look for the changes on imaging. T
(A) and ﬂuid-attenuated inversion recovery (FLAIR) (B) images show a large area
of hyperintensity in the right frontal region extending to the sylvian ﬁssure. Functional anisotropy (FA) map (C) shows a small area of low
FA at the right frontotemporal region posterior to the FLAIR abnormality. Abnormal to contralateral normal region FA and mean diffusivity
(MD) ratios showed 0.73470 and 1.06847, respectively. Color-coded FA fused with apparent diffusion coefﬁcient (ADC) map (D) shows
the abnormality more clearly. The cut-off value for the color-coded FA is kept at 0.2.
that in those without epilepsy. This suggests that DTI could
predict epileptogenesis in chronic TBI patients on the ba-
sis of reduced FA ratio. In the absence of histopathology
in the current studies, the reason for the observed reduc-
tion in the FA values is not known. However, based on
the literature reports, it appears that gliosis, at least in
part, could explain the reduced FA value. It was observed
in the earlier studies that the extent of gliosis correlates
with the epileptogenesis in patients with PTE (10). Glio-
sis, based on DTI and histopathology, also was implicated
in refractory epilepsy (30). Histopathologic studies of sur-
gically resected epileptogenic areas, which appeared nor-
mal on conventional MRI, have shown features of mild
white matter gliosis (41). Clusters of neuronal aggregates
in white matter and microdysgenesis (42) that could dis-
rupt the white matter tracts are thought to be responsible
for reduction in anisotropy in patients with epilepsy and
in malformations of the cortical development (28).
The increased MD ratio in TBI patients compared with
controls further supports a larger area of gliosis in chronic
TBI patients. This increased MD could result from a
number of pathologic mechanisms. Possible mechanisms
Epilepsia, Vol. 46, No. 9, 2005
DTI IN LATE POSTTRAUMATIC EPILEPSY 1469
FIG. 3. A 32-year-old man had a closed head injury about 2.5 years before imaging and was ﬁrst seen with seizures. T
(A) through the lateral ventricles shows encephalomalacia in the left frontal region with gliosis more prominently visible on ﬂuid-attenuated
inversion recovery (FLAIR) (B) imaging. Regions of interest (ROIs) placed on the fractional anisotropy (FA) map (C) posterior to the
abnormality and the corresponding contralateral region shows low FA (0.48739) and high mean diffusivity (MD) ratios (1.17008). Note the
large focal area with very low FA on the left side. Color-coded FA map fused with apparent. diffusion coefﬁcient (ADC) map (D) further
conﬁrms the large region in the left lobe that appears normal on T
and FLAIR images. Note the freehand ROIs placed in normal-appearing
region in the vicinity of the visible abnormality and the contralateral normal-appearing region (E) to calculate FA and MD ratios.
include reduced cell density and increased extracellular
space due to failure of neurogenesis or glial cell loss. In
a study in blunt head trauma, findings of increased mean
diffusivity suggested expansion of the extracellular space,
caused by neuronal or glial cell loss, which was not iden-
tified by conventional MRI despite the presence of neu-
rologic and neuropsychological symptoms and signs (6).
In our study, although the regional MD ratio was higher
in the TBI with epilepsy compared with the TBI without
epilepsy, it did not reach the level of statistical signifi-
cance. This is in keeping with a study by Wieshmann et al.
(43) in patients with partial epilepsy and structural abnor-
malities in which these authors demonstrated a reduced
anisotropy associated with a normal MD in 30% of pa-
tients with brain damage and dysgenesis. They attributed
these observations to the loss of directional organization
in which cell density was preserved.
Our findings of significantly decreased FA and the trend
of increased MD indicate that anisotropy may be a more
sensitive index than diffusivity in detecting the microstruc-
TABLE 1. Summary of regional mean FA and MD ratio in chronic TBI patients with and without epilepsy and controls
Patient Regional Regional p Value for p Value for
description No. FA ratio MD ratio FA ratio MD ratio
a. Controls 11 0.90 ± 0.067 0.95 ± 0.027 a vs. b (p = 0.0001) a vs. b (p = 0.006)
avs.c(p= 0.0001) a vs. c (p = 0.0001)
avs.d(p= 0.0001) a vs. d (p = 0.0001)
avs.e(p= 0.0001) a vs. e (p = 0.0001)
b. TBI without epilepsy from
/FLAIR normal region
9 0.68 ± 0.039 1.09 ± 0.141 b vs. c (p = 0.0001) b vs. c (p = 0.274)
c. TBI with epilepsy from T
14 0.57 ± 0.059 1.15 ± 0.140
d. TBI without epilepsy from
/FLAIR abnormal region
9 0.57 ± 0.203 2.02 ± 0.53
e. TBI with epilepsy from T
14 0.45 ± 0.131 2.54 ± 0.438 d vs. e (p = 0.123) d vs. e (p = 0.020)
FA, fractional anisotropy; MD, mean diffusivity; TBI, traumatic brain injury; No., number of patients; FLAIR, fluid-attenuated inversion recovery.
tural damage in chronic TBI patients. This observation
may be best explained by the concept that despite the dis-
organization of the tissue that causes decreased FA, the
cellular density is preserved in many of the chronic TBI
patients with epilepsy compared with that in those without
epilepsy. The pattern of changes is similar to that seen in
patients with brain damage and dysgenesis, commonly as-
sociated with epilepsy, in which anisotropy changes were
more evident than abnormalities of diffusivity in the de-
tection of structural cerebral lesions (43). This suggests
that anisotropy and diffusivity are diffusion entities that
give complementary information. However, the pattern of
changes are different in delayed PTE compared with that
observed in TLE patients, where diffusivity measurements
have been more sensitive than anisotropy in identifying the
DTI has been helpful in assessing the extent of the struc-
tural changes at and around the site of trauma. Rugg-Gunn
et al. (6) presented two cases of chronic TBI in which dif-
fuse axonal injuries distant from the major injury site were
present on DTI that were not visible on conventional MRI
Epilepsia, Vol. 46, No. 9, 2005
1470 R. K. GUPTA ET AL.
(6). In another case report, the patient’s excellent motor
recovery by 18 months after TBI correlated well with the
preservation of normal anisotropy in a portion of the pos-
terior limb of the internal capsule, an indication that the
pathways were intact (25). Huisman et al. (44) suggested
that changes in the diffusion anisotropy along the white
matter tracts might be of value in evaluating axonal dam-
age and can be used as a biomarker of TBI severity. In the
present study, extent of the damage as detected by using
DTI was larger than what was observed on conventional
MRI in all the patients with chronic TBI.
DTI provides information about the neuronal and mi-
crostructural organization of the epileptogenic lesion and
the surrounding tissue not observed with conventional
MRI techniques. Reduced diffusion anisotropy has re-
vealed abnormalities in patients with TLE (31), because
of decreased restriction on diffusion in directions perpen-
dicular to the axons. Reduced anisotropy and increased
diffusivity in the sclerotic hippocampi of patients with
chronic epilepsy and hippocampal sclerosis (45,46) sug-
gest the loss of structural organization and expansion of
the extracellular space. These studies indicate a potential
role for DTI in localizing the seizure focus in TLE. Surgi-
cal results are known to be poor in neocortical PTE (47).
The surgical results depend on the excision of the gliotic
region, as suggested by T
- and T
-weighted images and
localization of the abnormal region on intraoperative map-
ping with electrodes (48). It is believed that the incomplete
identification of the gliotic region is responsible for failure
of surgery in neocortical PTE (47). The finding of signif-
icantly altered anisotropy within normal-appearing tissue
could be of potential clinical importance in detecting oc-
cult epileptogenic regions and may help in offering better
surgical results in these patients.
In the current studies, we used the ipsilateral and con-
tralateral ratios of FA and MD for determining the relative
changes in DTI parameters for identifying the abnormal
regions beyond the T
/FLAIR lesions. However, these ra-
tios could mask some of the pathology in the presence of
The determination of the DTI metrics was based on
the ROI analysis. The ROI analysis has both advantages
and disadvantages. When the placement of the ROI was
done by an expert neuroradiologist, as was done in these
studies, the ROI analysis can yield accurate and repro-
ducible results about individual patients. However, from
a practical point of view, the ROI analysis is limited to a
fewer regions and can introduce possible bias into the ROI
selection. An alternative approach is to perform voxel-
based morphometry (49), in which a single patient datum
can be compared with a group of controls. The advan-
tage of such an approach is that differences between two
groups can be determined with high spatial resolution in
an unbiased manner. Such an approach was adapted by
Rugg-Gunn et al. (30) in the DTI analysis of refractory
epilepsy. However, as pointed out by Bookstein (50), and
more recently by Davtzikos (51), voxel-based morphom-
etry also has a few disadvantages (50,51). For instance,
accurate spatial normalization is critical for voxel-based
morphometry to work well. Accurate spatial registration
can be realized if the brain has only subtle pathology with-
out major anatomic distortions. This is not always the case
with TBI patients. Thus when brain is affected by diseases,
caution must be exercised when drawing inferences based
on voxel-based morphometry.
These DTI studies demonstrate the existence of signifi-
cant pathology beyond the obvious lesions seen on the T
and FLAIR images in chronic TBI. The extent and sever-
ity of the DTI-observed pathology may help in separating
PTE from those without epilepsy in chronic TBI patients
on the basis of reduced regional FA ratio. These findings
could improve the understanding of the pathophysiology
of epilepsy and in planning specific surgical strategies re-
lated to PTE.
1. Kurtzke JF, Kurland LT, The epidemiology of neurologic disease.
In:Baker AB, Joynt RJ, eds. Clinical neurology.New York: Harper
2. Hernandez TD, Naritoku DK. Seizures, epilepsy and functional
recovery after traumatic brain injury: a reaappraisal. Neurology
3. Dunn LT, Foy PM. Anticonvulsant and antibiotic prophylaxis in
head injury. Ann R Coll Surg Engl 1994;76:147–9.
4. Lee ST, Lui TN, Wong CW, et al. Early seizures after moderate
closed head injury. Acta Neurochir (Wien) 1995;137:151–4.
5. Caveness WF. Epilepsy, a product of trauma in our time. Epilepsia
6. Rugg-Gunn FJ, Symms MR, Barker GJ, et al. Diffusion imaging
shows abnormalities after blunt head trauma when conventional
magnetic resonance imaging is normal. J Neurol Neurosurg Psy-
7. Kimura H, Meaney DF, McGowan JC, et al. Magnetization trans-
fer imaging of diffuse axonal injury following experimental brain
injury in the pig: characterization by magnetization transfer ra-
tio with histopathological correlation. J Comput Assist Tomogr
8. Bagley LJ, McGowan JC, Grossman RI, et al. Magnetization trans-
fer imaging of traumatic brain injury. JMagn Reson Imaging
9. Crooks DA. The pathological concept of diffuse axonal injury: its
pathogenesis and the assessment of severity. JPathol 1991;165:5–
10. Kumar R, Gupta RK, Husain M, et al. Magnetization transfer MR
imaging in patients with posttraumatic epilepsy. AJNR Am J Neu-
11. Moseley ME, Cohen Y, Mintorovitch J, et al. Early detection of
regional cerebral ischemia in cats: comparison of diffusion- and T
weighted MRI and spectroscopy. Magn Reson Med 1990;14:330–
12. Tien RD, Felsberg GJ, Friedman H, et al. MR imaging of high-grade
cerebral gliomas: value of diffusion-weighted echoplanar pulse se-
quences. AJR Am J Roentgenol 1994;162:671–7.
13. Warach S, Gaa J, Siewert B, et al. Acute human stroke studied
by whole brain echo planar diffusion-weighted magnetic resonance
imaging. Ann Neurol 1995;37:231–41.
14. Droogan AG, Clark CA, Werring DJ, et al. Comparison of multi-
ple sclerosis clinical subgroups using navigated spin echo diffusion
weighted imaging. Magn Reson Imaging 1999;17:653–61.
Epilepsia, Vol. 46, No. 9, 2005
DTI IN LATE POSTTRAUMATIC EPILEPSY 1471
15. Liu AY, Maldjian JA, Bagley LJ, et al. Traumatic brain injury:
diffusion-weighted MR imaging findings. AJNR Am J Neuroradiol
16. Le Bihan D. Looking into the functional architecture of the brain
with diffusion MRI. Nat Rev Neurosci 2003;4:469–80.
17. Basser PJ, Mattiello J, Le Bihan D. MR diffusion tensor spec-
troscopy and imaging. Biophys J 1994;66:259–67.
18. Basser PJ, Mattiello J, Le Bihan D. Estimation of the effec-
tive self-diffusion tensor from NMR spin echo. JMagn Reson B
19. Pierpaoli C, Jezzard P, Baser PJ, et al. Diffusion tensor MR imaging
of the human brain. Radiology 1996;201:637–48.
20. Basser PJ, Pierpaoli C. Microstructural and physiological features
of tissues elucidated by quantitative-diffusion-tensor MRI. JMagn
Reson B 1996;111:209–19.
21. Le Bihan D, Mangin JF, Poupon C, et al. Diffusion tensor imaging
concepts and applications. JMagn Reson Imaging 2001;13:534–
22. Alsop DC, Connelly A, Duncan JS, et al. Diffusion and perfusion
MRI in epilepsy. Epilepsia 2002;43(suppl):69–77.
23. van Gelderen P, de Vleeschouwer MHM, DesPres D, et al. Water
diffusion and acute stroke. Magn Reson Med 1994;31:154–63.
24. Lythgoe MF, Busza AL, Calamante F, et al. Effects of diffusion
anisotropy on lesion delineation in a rat model of cerebral ischemia.
Magn Reson Med 1997;38:662–8.
25. Werring DJ, Clark CA, Barker GJ, et al. The structural and functional
mechanisms of motor recovery: complementary use of diffusion ten-
sor and functional magnetic resonance imaging in a traumatic injury
of the internal capsule. J Neurol Neurosurg Psychiatry 1998;65:863–
26. Lim KO, Hedehus M, Moseley M, et al. Compromised white mat-
ter tract integrity in schizophrenia inferred from diffusion tensor
imaging. Arch Gen Psychiatry 1999;56:367–74.
27. Werring DJ, Clark CA, Barker GJ, et al. Diffusion tensor imaging
of lesions and normal-appearing white matter in multiple sclerosis.
28. Eriksson SH, Rugg-Gunn FJ, Symms MR, et al. Diffusion tensor
imaging in patients with epilepsy and malformations of cortical de-
velopment. Brain 2001;124:617–26.
29. Rugg-Gunn FJ, Eriksson SH, Symms MR, et al. Diffusion ten-
sor imaging of cryptogenic and acquired partial epilepsies. Brain
30. Rugg-Gunn FJ, Eriksson SH, Symms MR, et al. Diffusion tensor
imaging in refractory epilepsy. Lancet 2002;359:1748–51.
31. Arfanakis K, Hermann BP, Rogers BP, et al. Diffusion tensor MRI
in temporal lobe epilepsy. Magn Reson Imaging. 2002;20:511–9.
32. Basser PJ. Inferring microstructural features and the physiologi-
cal state of tissues from diffusion-weighted images. NMR Biomed
33. Conturo TE, McKinstry RC, Aronovitz JA, et al. Diffusion MRI:
precision, accuracy and flow effects. NMR Biomed 1995;8:307–
34. Reese TG, Heid O, Weisskoff RM, et al. Reduction of eddy-current-
induced distortion in diffusion MRI using a twice-refocused spin
echo. Magn Reson Med 2003;49:177–82.
35. Hasan KM, Parker DL, Alexander AL. Comparison of gradient en-
coding schemes for diffusion-tensor MRI. JMagn Reson Imaging
36. Batchelor PG, Atkinson D, Hill DL, et al. Anisotropic noise propa-
gation in diffusion tensor MRI sampling schemes. Magn Reson Med
37. Hasan KM, Narayana PA. Computation of the fractional anisotropy
and mean diffusivity maps without tensor decoding and diago-
nalization: theoretical analysis and validation. Magn Reson Med
38. Hasan KM, Alexander AL, Narayana PA. Does fractional anisotropy
have better noise immunity characteristics than relative anisotropy
in diffusion tensor MRI? An analytical approach. Magn Reson Med
39. Woods RP, Grafton ST, Holmes CJ, et al. Automated image registra-
tion, I: general methods and intrasubject, intramodality validation.
J Comput Assist Tomogr 1998;22:139–52.
40. Jennet WB. Epilepsy after non-missile head injuries. 2nd ed. Lon-
don: William Heinemann Medical Books, 1975.
41. Palmini A, Andermann F, Olivier A, et al. Focal neuronal migra-
tion disorders and intractable partial epilepsy: a study of 30 pa-
tients.[Review]. Ann Neurol 1991;30:741–9.
42. Hardiman O, Burke T, Phillips J, et al. Microdysgenesis in re-
sected temporal neocortex: incidence and clinical significance in
focal epilepsy. Neurology 1988;38:1041–7.
43. Wieshmann UC, Clark CA, Symms MR, et al. Reduced anisotropy
of water diffusion in structural cerebral abnormalities demonstrated
with diffusion tensor imaging. Magn Reson Imaging 1999;17:1269–
44. Huisman TAGM, Schwamm LH, Schaefer PW, et al. Diffusion ten-
sor imaging as potential biomarker of white matter injury in diffuse
axonal injury. AJNR Am J Neuroradiol 2004;25:370–6.
45. Wieshmann U, Clark C, Symms M, et al. Water diffusion in the
human hippocampus in epilepsy. Magn Reson Imaging 1999;17:29–
46. Yoo S, Chang K, Song I, et al. Apparent diffusion coefficient value
of the hippocampus in patients with hippocampal sclerosis and in
healthy volunteers. AJNR Am J Neuroradiol 2002;23:809–12.
47. Marks DA, Kim J, Spencer DD, et al. Seizure localization and pathol-
ogy following head injury in patients with uncontrolled epilepsy.
48. Kazemi NJ, So EL, Mosewich RK, et al. Resection of frontal en-
cephalomalacias for intractable epilepsy: outcome and prognostic
factors. Epilepsia 1997;38:670–7.
49. Ashburner J, Friston KJ. Voxel-based morphometry: the methods.
50. Bookstein FL. “Voxel-based morphometry” should not be used with
imperfectly registered images. NeuroImage 2001;14:1454–62.
51. Davatzikos C. Why voxel-based morphometric analysis should be
used with great caution when characterizing group differences. Neu-
Epilepsia, Vol. 46, No. 9, 2005