Evaluation of Functional MRI and DTI (Imaging Techniques) in Children With Epilepsy and Focal Brain Lesions

Introduction

Background and Significance • Functional MRI Surgical treatment is being used with
increasing frequency for patients with intractable epilepsy. The success of this operation
largely depends on the results of a comprehensive pre-operative patient evaluation with the
main purpose of delineating the epileptogenic lesion. The pre-operative assessment includes
video EEG monitoring, structural neuroimaging (i.e. MRI), functional cerebral imaging such
as SPECT (Single Photon Emission Computed Tomography) and PET (Positron Emission Tomography)
which are capable of localizing epileptogenic foci, and neuropsychological evaluation. The
more results of these localization techniques that coincide to a single epileptogenic focus,
the greater the likelihood for a successful surgical outcome. In the preoperative evaluation
of surgical candidates, a unique neuropsychological evaluation called the Wada test is
performed. The Wada test includes standard neuropsychological assessment and intracarotid
sodium amytal testing, to have the capability of predicting hemispheric lateralization of
language and memory and sometimes helping with localization of a brain lesion based on
cognitive function and dysfunction [1]. The Wada test is usually used to determine the
cerebral speech dominance, to predict postsurgical amnesia and is found to be useful in
predicting laterality of seizure focus in candidates for temporal lobectomy. However, the
Wada test is invasive, requiring an angiogram. In the mean time, functional magnetic
resonance imaging (fMRI) has been proven to have potential as a possible non-invasive
alternative method for lateralizing the language and motor networks, and to a lesser degree
– memory, in cooperative children with epilepsy [2-4]. Functional MRI also provides an
opportunity to follow the post-lesional or post-surgery plasticity of these networks using
repeated examinations in a same patient, such that fMRI may also contribute additional
useful information in the pre-surgical planning and post-surgical monitoring. Therefore, the
effectiveness and the reliability of fMRI in epilepsy patients compared to Wada test (in
those patients where the Wada test is deemed clinically relevant) would be very helpful in
assessing the clinical utility of fMRI in epilepsy patients.

Another patient population where presurgical MR imaging evaluation is performed are children
with focal brain lesions. It is important to know where eloquent cortex lies in relationship
to these abnormalities as well as identifying important anatomic structures. Performing
advanced imaging evaluation of pediatric neurosurgical candidates provides greater insight
to the nature of the abnormality that is present and offers guidance for the neurosurgeon in
the removal of the lesion as safely as possible.

A number of studies have shown that fMRI language tasks reliably identify language areas in
presurgical patients, but activation using single paradigms sometimes may disagree with the
Wada test [5-9]. For instance, most studies report partial disparity with Wada test in 10
to 15% of studies [10]. Therefore, the hope for fMRI as a means to replace Wada test for
language lateralization is hampered. A likely source of those occasional discrepancies is
that language activation tasks, for both Wada test and fMRI, target only the limited aspects
of language processing. Using two or more tasks could provide a better representative view
of language laterality by mapping multiple language functions [10]]. By doing that, we
should be able to ascertain whether a series of tasks confirm findings, improve rater
reliability, and increase agreement with Wada test (in those patients where Wada is deemed
clinically relevant). In this study, we will employ tasks designed to identify anterior
“expressive” networks in inferior frontal gyrus (IFG), implicated in word retrieval, and
middle frontal gyrus (MFG), implicated in verbal working memory. We also will use tasks to
identify “receptive” areas along the left superior temporal sulcus implicated in
whole-language comprehension. Furthermore, tasks were designed so that at least two tasks
targeted similar language/processing functions and regions.

• Diffusion Tensor Imaging The second imaging technique to be employed in this project is
diffusion tensor imaging, which provides a quantitative measure of the microstructure of
organized tissue as well as the direction of this organization. In organized tissue such as
brain white matter and muscle, diffusion of water is orientation dependent. In the past
decade, the development of diffusion tensor imaging with MRI has made it possible to
quantitatively map this anisotropy noninvasively, providing a unique means for studying
tissue orientation and microstructural integrity in vivo [11-13]. Although there are a
number of possible contributions to the diffusion anisotropy [14], its measurement has been
applied to many in vivo applications because of it being now well established that the MR
measurement of the effective diffusion tensor in tissues can provide unique biologically and
clinically relevant information that cannot be provided by other imaging methods. The most
interesting application of DTI lies in the study of brain white matter.

Because of the sensitivity of DTI in detecting degradations of microstructural integrity of
white matter, it is natural to apply DTI to the study of neurological pathways of language
networks in the patients with intractable epilepsy. The focus of the DTI work in this study
will be to evaluate the white matter of neural circuits serving the regulation of language
processing and compare it with that in the region of counter lateral side of the brain.
Thus, the areas to be focused on in this study will include white matter tracts between
Broca’s and Wernicke’s areas on the dominant side of language compared to the counter
lateral side. The analysis of DTI data mostly relied on measurements in ROIs defined using a
priori knowledge. DTI will also be used to identify major white matter tracts in order to
evaluate their involvement by the abnormality being investigated and to provide a map for
the neurosurgical approach for lesion resection.

Future participation as one of the sites in the multi-center trial for fMRI Mapping in
Childhood Epilepsy, lead by William Gaillard, M.D., P.I. at Children’s National Medical
Center is anticipated. The data for the fMRI portion proposed in this study will be shared
with the other sites participating in the multi-center trial by complying with the federal
regulations for data sharing.

Objectives:

Objective 1. Investigate the responses of brain functions to fMRI tasks targeting different
aspects of language processing to ascertain whether these tasks confirm findings and result
in agreement with Wada test (in those patients where the Wada test is deemed clinically
relevant). Our hypothesis is that a panel of fMRI tasks targeting different aspects of
language processing increases accuracy in determining hemisphere language dominance.

Objective 2. Examine the relationship between the specific language deficits and brain
morphology. To examine these relationships, we will utilize an imaging protocol that will
include evaluation of brain structure and connectivity in white matter tracts using
diffusion tensor imaging (DTI).

Objective 3. Investigate the responses of brain functions to fMRI tasks stimulated by motor
left/right finger tapping to assist the evaluation of the geometric relationship between
left/right motor cortex and seizure foci identified by SPECT. The hypothesis is that the
result of this evaluation should be helpful in surgical planning for the epilepsy patients
and focal brain lesion neurosurgical candidates.

Methods and Materials

Human Subjects: 30 patients with intractable epilepsy and/or focal brain lesions and up to
20 normal volunteers, ranging in age from 4 to 21 years, will participate with fMRI language
paradigms, fMRI motor mapping, and diffusion tensor imaging (DTI). Patients requiring
sedation for scanning will be excluded from the study.

Prior to the evaluations summarized in this protocol, the patients recruited for this study
will undergo the regular pre-surgical neuropsychological tests, and the Wada test where
appropriate. The Wada test will be limited to those patients where the test is deemed
clinically relevant and ordered as standard of care.

Neuroimaging:

The MRI experiments will be performed on a 1.5T Siemens scanner located at CHOA’s Scottish
Rite campus equipped with an array head coil. Anatomical and DTI sequences are already
established on this scanner and preliminary fMRI data has been acquired. Fast imaging
sequences, using echoplanar methods, will be used to enable the high temporal resolution
necessary for fMRI. In addition to the functional scans, a series of conventional anatomic
scans will be obtained for each participant, which will allow the subsequent superposition
of functional data with anatomy. The scan will include structural MRI, fMRI while a series
of cognitive (language) tasks and motor tasks are performed, and then the DTI. All scans
should be completed in 30-60 minutes.

Functional MRI: fMRI images will be acquired using a single-shot T2*-weighted EPI. During
the gradient EPI scanning, the stimuli will be used to activate the particular brain
function. A multi-task stimulation paradigms using a block design approach, including
listening, reading, semantic decision, and verbal fluency tasks, will be applied in order to
increase the confidence level of the lateralization of the language function. The tasks
include a response element and the response is monitored using a button box. For all the
tasks, the subjects do not talk during MR scanning to avoid motion artifacts. When a task
requires reading or answering, the subjects should do it silently without moving head or
mouse as best as they can. The subjects understanding of the task will be validated by
analysing the responses from the button box. For those children with epilepsy selected for
this study, they will be divided into four age groups (4-6, 7-9, 10-12, 12+). The
difficulty level of the tasks will be selected before the experiment, which is
proportionally corresponding to the age group that the child belongs.

Task 1. Listening to a story Task 2. Reading a Story Task 3. Auditory description task (ADT)
– Semantic Decision Task 4. Auditory Category Task (Autcat or AC) – Verbal Fluency

For the motor tasks a series of cognitive tasks (left and/or right finger tapping) will be
performed.

Diffusion Tensor Imaging: DTI is becoming a widely available and used technique for
studying white matter integrity in the brain. While it is possible to perform tract-tracing
based on diffusion tensor data, results are still not adequately reliable for routine
applications. Thus, our approach will utilize mainly the magnitude of anisotropy,
particularly fractional anisotropy in this study. Many studies (e.g. (Lim and Helpern
2002)) using anisotropy measurement to ascertain microstructural degradation in various
neurological and psychiatric diseases have been reported. In this study, we will perform
the DTI imaging of the whole brain and use ROI-based analysis of fractional anisotropy to
ascertain differences in white matter integrity about language networks between different
subjects with intractable epilepsy. Using this DTI method we will also identify major white
matter fiber tracks that are important to visualize for neurosurgical planning.

The diffusion tensor images will be acquired with a single-shot EPI sequence with diffusion
weighting applied in 12 directions. The DTI data set will be transferred onto a workstation
where software for DTI post-processing is installed for diffusion tensor calculation on a
pixel-by-pixel basis. Based on the diffusion tensor information, fractional anisotropy (FA)
maps will be generated to reveal the degree of directional dependence of the diffusion in
brain tissue (FA=1 when diffusion is purely anisotropic and FA=0 when diffusion is purely
isotropic) and maps of the apparent diffusion coefficient (ADC) will also be calculated. In
order to visualize white matter tracts, the principal eigenvector associating with the
maximum eigenvalue will be used to obtain the projections along x, y, z directions in lab
frame, respectively, which will be color coded and displayed to represent the orientation of
white matter tracts. ROIs along the language networks/pathways will be defined based on the
anatomical knowledge as well as the results from functional connectivity study. Average FA
and ADC in the ROIs will be calculated and compared with the counter lateral side of the
brain and with different subject in the analysis.

Data Analysis: The image data will be transferred to a PC in a locked room and will require
a password to access the data in order to protect patients’ confidentiality. The image data
will be analyzed using the software installed on PC to obtain the fMRI and DTI results. The
original image data will also be encoded and de-identified, and then sent to Central Storage
and Processing to be accessed by all sites who participate in the multi-center trial as
referenced earlier.

Potential risks and discomforts: Potential risks associated with this participation in the
research project are minimal.

Imaging procedures. There are four parameters that are considered as potential risks in a MR
study: (a) static magnetic field strength, (b) rate of change of magnetic field (dB/dt), (c)
RF power desposition, and (d) acoustic noise level. On the 1.5 T clinical systems which will
be used for these studies, all of these parameters are kept within the FDA limits and do not
pose a significant risks.

As mentioned above, there are no known risks associated with the MRI procedure although
there may be some risks that are not known. The magnet makes a loud pinging sound; however,
earphones are worn during the procedure to muffle this sound and to allow the technician to
communicate with the participant during the procedure. Because the magnetic field affects
any metallic object, participants cannot be included if they have any type of metallic
implant in the body, including pacemakers, aneurysm clips, shrapnel, metal fragments,
orthopedic pins, screws or plates, IUDs or body piercings that cannot be removed. The risks
associated with these objects (that they may heat up or move) are discussed with the
potential participants when they are screened.

There are also minor discomforts encountered by some participants during the procedure,
including muscle discomfort from lying still in the scanner, or becoming too hot or cold due
to the ambient temperature of the room. Some people become claustrophobic in the scanner.
Others experience a feeling of dizziness in response to the strong magnetic field (see
above). These possibilities are explained to participants and they are requested to ask the
staff for help if they are uncomfortable. Participants are informed that they are free to
stop the procedure at any time.

All the medical tests and neuropsychological procedures pre- or post- imaging studies
involved in this research project are regular routine medical procedures that will be
conducted at Children’s Healthcare of Atlanta at Scottish Rite. The questionnaires before
the imaging session do not involve risk to the participants.

Schedule and Procedures:

A total of 30 patients and up to 20 normal volunteers will participate this research study.
Before the imaging session, a screening questionnaire will be administered and an informed
consent/assent procedure will be completed. Participants in the “normal volunteer” group
will be reimbursed $50 for their time and other expenses associated with participation. The
schedule of procedures for the neuroimaging is as follows:

1. Transportation. The participants should provide their own transportation to Children’s
Healthcare of Atlanta at Scottish Rite where the MRI scan will take place.

2. Task training. Before the imaging procedures, participants will complete a training
session on the tasks they will be asked to complete while in the scanner. This is a
computer-based training program; training to criterion should take less than one hour.
When training is complete, the participants will be taken to the MRI suite for the MRI
scan.

3. Head Scanning: The structural magnetic resonance imaging (sMRI), functional magnetic
resonance imaging (fMRI) and diffusion tensor imaging (DTI) will be performed on a
Siemens 1.5T clinical scanner at Children’s Healthcare of Atlanta at Scottish Rite.
An array head coil will be used to collect the MR data. Fast imaging sequences, using
echoplanar methods, will be used to enable the high temporal resolution necessary for
fMRI. A series of conventional anatomic scans will be obtained for each participant,
which will allow the subsequent superposition of functional data with anatomy.

4. Tasks: While the fMRI is being performed, the participants will be asked to perform a
series of cognitive tasks assessing language and motor function. Audio and visual
stimuli will be presented and participants will respond by pressing a button. Images
will be projected on a PC interfaced to a long focal length projector. These images
will thus be projected into the magnet bore and will be visible to the participant via
a small mirror on the head holder. Responses will be acquired with a fiber optic
button response box interface to the computer. The fiber optic box allows response
acquisition without the generation of electromagnetic noise in the scanner. All
scanning will be completed in 30-60 minutes,