168
K. Miller et al.
A number of groups have investigated the use of intra-operative ultrasound in
order to provide data that could be used to update pre-operative models to account
for brain shift [20, 21]. Laser range scanners (LRS) [13, 22] and stereo cameras
[22] have also been used. However the data obtained using these techniques cannot
be directly used for surgical navigation, as images typically have poor contrast and
resolution, and surface data alone is not sufficient to guide neuronavigation.
Summary
1. Inaccuracies in the identification of invasive electrode placements and the limited
ability to relate the SOZ identified by these electrodes to features of brain
anatomy prevent iEEG data from being fully utilised to identify appropriate
resection regions.
2. A method to reliably and precisely identify the location of seizure-onset zones
could lead to curing millions of patients of epilepsy.
3. The key technology we describe in this chapter is biomechanics-based prediction
of brain deformations resulting from the invasive electrode placement and the
use of these deformations to register a pre-operative MRI image onto an intraoperative CT image with electrodes in place (Fig. 7.1d).
In Sect. 7.2 we describe how to adapt the methods presented in Chap. 6 to
model brain deformations resulting from invasive electrode placement. In Sect.
7.3, to demonstrate the appropriateness and effectiveness of our biomechanicsbased methods, we consider results of brain deformation computation for a specific
example of a paediatric epilepsy patient from Boston Children’s Hospital. The
numerical algorithms devised to efficiently solve brain deformation models are
described in Chaps. 10 and 11.
7.2 Computing Brain Deformations Due to Insertion
of Invasive Electrodes
Biomechanical modelling issues related to geometry and finite element mesh,
boundary conditions, loading and material properties are briefly summarised below.
For a more detailed discussion of these issues, please refer to Chap. 6.
A high-level depiction of the inputs, outputs and software packages used in each
step of the analysis is shown in Fig. 7.2, next page.
7.2.1 Geometry
The first step in the construction of model geometry is to develop surface models
of the brain parenchyma and ventricles. 3D Slicer (http://www.slicer.org/) is used
to segment the regions of interest (ROI) from the patient’s MRI with the help of
K. Miller et al.
A number of groups have investigated the use of intra-operative ultrasound in
order to provide data that could be used to update pre-operative models to account
for brain shift [20, 21]. Laser range scanners (LRS) [13, 22] and stereo cameras
[22] have also been used. However the data obtained using these techniques cannot
be directly used for surgical navigation, as images typically have poor contrast and
resolution, and surface data alone is not sufficient to guide neuronavigation.
Summary
1. Inaccuracies in the identification of invasive electrode placements and the limited
ability to relate the SOZ identified by these electrodes to features of brain
anatomy prevent iEEG data from being fully utilised to identify appropriate
resection regions.
2. A method to reliably and precisely identify the location of seizure-onset zones
could lead to curing millions of patients of epilepsy.
3. The key technology we describe in this chapter is biomechanics-based prediction
of brain deformations resulting from the invasive electrode placement and the
use of these deformations to register a pre-operative MRI image onto an intraoperative CT image with electrodes in place (Fig. 7.1d).
In Sect. 7.2 we describe how to adapt the methods presented in Chap. 6 to
model brain deformations resulting from invasive electrode placement. In Sect.
7.3, to demonstrate the appropriateness and effectiveness of our biomechanicsbased methods, we consider results of brain deformation computation for a specific
example of a paediatric epilepsy patient from Boston Children’s Hospital. The
numerical algorithms devised to efficiently solve brain deformation models are
described in Chaps. 10 and 11.
7.2 Computing Brain Deformations Due to Insertion
of Invasive Electrodes
Biomechanical modelling issues related to geometry and finite element mesh,
boundary conditions, loading and material properties are briefly summarised below.
For a more detailed discussion of these issues, please refer to Chap. 6.
A high-level depiction of the inputs, outputs and software packages used in each
step of the analysis is shown in Fig. 7.2, next page.
7.2.1 Geometry
The first step in the construction of model geometry is to develop surface models
of the brain parenchyma and ventricles. 3D Slicer (http://www.slicer.org/) is used
to segment the regions of interest (ROI) from the patient’s MRI with the help of