Direct Head Tracking - Robotized Transcranial Magnetic Stimulation - page 142

8.4 Head Contour Generation Based on Laser Scans
For robotized and neuro-navigated TMS, a 3D contour of the patient's head is
required as navigation source. The coil pose is then calculated in relation to the
contour. Furthermore, the stimulation targets are planned and documented by
means of the contour. Conventionally, an MRI scan of the patient is performed
prior to stimulation. On the basis of the MRI scan the contour is then generated.
However, MRI scanning time is always short in a clinical setting and expensive.
For treatment attempts using repetitive Transcranial Magnetic Stimulation (rTMS)
or brain mapping purposes the underlying brain structure is essential for successful
and meaningful investigations. For many other TMS applications the underlying
brain topology is not necessarily required: For research applications and investi-
gations, a hot spot search is performed finding an optimal stimulation point for
each single subject by measuring the Motor Evoked Potentials (MEPs) of a spe-
cific muscle. The underlying brain structure is not needed for this purpose [ 1 ].
Instead, it is more important to stimulate precisely at the hot spot and to re-access
the hot spot in different trials or days.
For TMS experiments with the described setup, the patient's head is commonly
registered to a standard head. For instance, the head contour of a head phantom is
frequently used as reference standard head. However, this standard head can only
be used as a coarse approximation. As each human head differs in size and shape,
the difference between real head and used standard head can be large. This can
lead to systematic misalignments of the coil when using the robotized TMS system
as the robot orientates the coil tangentially by means of the contour. To overcome
that, we propose to use a 3D laser scanning system to obtain an individual 3D
contour of the subject's head.
8.4.1 Head Scanning and Contour Generation
In accordance with the setup for direct head tracking with laser scans (see
Sect. 8.3.1 ), we use the GALAXY laser system for head scanning. We have
already shown that a high resolution scan consists of roughly 10 ; 000 surface
points and that a registration of these points to a MRI generated head contour can
be performed with a mean error of approximately 0.3 mm (cf. Sect. 8.3.3.2 ).
We now use this high resolution laser scan to generate a smooth head contour.
Therefore, we are applying the PowerCrust algorithm [ 2 ]. Figure 8.10 b shows the
generated contour of our head phantom. In Fig. 8.10 a the contour is illustrated
with the underlying data points. Hair is a critical issue as hair absorbs laser light.
Therefore, we use white swimming caps that are tight-fitting to the head. Note that
this cap is only needed for laser scan acquisition. In contrast, the MRI-based head
contour is generated using edge detection extracting the head surface in the MRI
images.