Clin Neuroradiol (2012) 22:55–68
DOI 10.1007/s00062-012-0136-3
O r i g i n a l A rt i c l e
Visualisation of the Zona Incerta for Deep  
Brain Stimulation at 3.0 Tesla
H. U. Kerl · L. Gerigk · S. Huck · M. Al-Zghloul ·
C. Groden · I. S. Nölte
Received: 5 September 2011 / Accepted: 27 January 2012 / Published online: 17 February 2012
© Springer-Verlag 2012
Abstract
Purpose  Deep-brain stimulation (DBS) of the zona incerta
(ZI) has shown promising results for medication-refractory
neurological disorders including Parkinson’s disease (PD)
and essential tremor (ET). The success of the intervention
is indispensably dependent on the reliable visualisation of
the ZI.
The aim of the study was to evaluate different promi-
sing new magnetic resonance imaging (MRI) methods at
3.0 Tesla for pre-stereotactic visualisation of the ZI using a
standard installation the protocol.
Methods  MRI of nine healthy volunteers was acquired
(T1-MPRAGE, T2-FLAIR, T2*-FLASH2D, T2-SPACE
and susceptibility-weighted imaging (SWI). Image quality
and visualisation of the ZI for each sequence were analysed
independently by two neuroradiologists using a 6-point
scale. For T2*-FLASH2D the axial, coronal and sagittal
planes were compared. The delineation of the ZI versus the
internal capsule, the subthalamic nucleus and the pallido-
fugal fibres was evaluated in all sequences and compared
to T2-FLAIR using a paired t-test. Inter-rater reliability,
contrast-to-noise ratios (CNR), and signal-to-noise ratios
(SNR) for the ZI were computed. For illustration, coronal
T2*-FLASH2D images were co-registered with the cor-
I. S. Nölte, MD () · H. U. Kerl, MD · S. Huck, MD ·
M. Al-Zghloul, MD · C. Groden, MD
Medical Faculty Mannheim,
Department of Neuroradiology, University of Heidelberg,
Theodor-Kutzer-Ufer 1–3, 68167 Mannheim, Germany
e-mail: ingo.noelte@umm.de
L. Gerigk, MD
Division of Radiology, German Cancer Research Centre,
Im Neuenheimer Feld 280, 69120 Heidelberg, Germany
responding section schema of the Schaltenbrand–Wahren
stereotactic atlas.
Results  Only the rostral part of the ZI (rZI) could be iden-
tified. The rZI was best and reliably visualised in T2*-
FLASH2D (particularly coronal orientation; p < 0.05). No
major artifacts in the rZI were observed in any of the se-
quences. SWI, T2-SPACE, and T2*-FLASH imaging of-
fered significant higher CNR values for the rZI compared
to T2-FLAIR imaging using standard parameters. The
co-registration of the coronal T2*-FLASH2D images pro-
jected the ZI clearly into the boundaries of the anatomical
sections.
Conclusions The delineation of the rZI is best possible in
T2*-FLASH2D (particularly coronal view) using a stand-
ard installation protocol at 3.0 T. The caudal ZI could not
be discerned in any of the sequences.
Keywords  Deep brain stimulation (DBS) ·
Essential tremor (ET) · Magnetic resonance imaging
(MRI) · Parkinson’s disease (PD) · Zona incerta (ZI)
Darstellung der Zona incerta bei 3.0 Tesla  
für die Tiefenhirnstimulation
Zusammenfassung
Zielsetzung  Die Tiefenhirnstimulation der Zona incerta
(ZI) stellt eine neue, Erfolg versprechende Therapieoption
für medikamenten-refraktäre neurologische Erkrankungen,
wie z. B. Morbus Parkinson und essentiellen Tremor, dar.
Der Erfolg der Methode ist entscheidend von einer ge-
nauen und verlässlichen präoperativen Darstellung der ZI
abhängig.
13
56
Das Ziel der Studie war es, verschiedene MRT-Methoden
für die präoperative Darstellung der ZI unter Verwendung
einer Standardkonfiguration bei 3.0 T zu evaluieren.
Methoden  Für die Studie wurden 3.0 T MRT-Bilder von
neun gesunden Probanden akquiriert (T1-MPRAGE, T2-
FLAIR, T2*-FLASH2D, T2-SPACE und suszeptibilitäts-
gewichtete Sequenzen (SWI)). Zwei Neuroradiologen
bewerteten unabhängig voneinander die Bildqualität und
die Abgrenzbarkeit der ZI für jede Sequenz. Für T2*-
FLASH2D Sequenzen erfolgte ein Vergleich der axialen,
koronaren und sagittalen Schichtung. Die Abgrenzbarkeit
der ZI gegenüber der inneren Kapsel, dem Nucleus sub-
thalamicus und den pallidofugalen Fasern wurde für alle
Sequenzen beurteilt und mittels eines paired t-test gegen-
über T2-FLAIR verglichen. Als quantitative Parameter
wurden das Signal-zu-Rausch-Verhältnis (SNR) sowie das
Kontrast-zu-Rausch-Verhältnis (CNR) für die ZI berech-
net. Zur Veranschaulichung wurde ein Bild der koronaren
T2*-FLASH2D Sequenz auf Höhe der ZI mit einer digi-
talisierten koronaren Schemazeichnung des Schaltenbrand-
Wahren Atlas für Stereotaxie co-registriert.
Ergebnisse  Während der caudale Anteil der ZI nicht abge-
grenzt werden konnte, war die genaueste und verlässlichste
Darstellung der rostralen ZI (rZI) mittels T2*-FLASH2D
(insbesondere den koronaren Akquisitionen; p < 0.05) mög-
lich. Die SWI, T2-SPACE und T2*-FLASH2D Sequen-
zen zeigten zudem einen signifikanten Anstieg der CNR
im Vergleich zu T2-FLAIR. In der Co-Registrierung der
koronaren T2*-FLASH2D Bilder und der digitalisierten
Schemazeichnung konnte die ZI eindeutig in ihren anato-
mischen Grenzen nachgewiesen werden.
Schlussfolgerung  Unter Verwendung einer Standardkonfi-
guration im 3.0 T MRT bietet T2*-FLASH2D (insbesonde-
re die koronare Akquisition) die derzeit bestmöglich prä-
operative Darstellung der rZI. Die caudale ZI konnte in
keiner der verwendeten Sequenzen abgegrenzt werden.
Schlüsselwörter  Essentieller Tremor ·
Magnetresonanztomographie · Morbus Parkinson ·
Tiefenhirnstimulation · Zona incerta
Introduction
Deep brain stimulation (DBS) is a reversible stereotac-
tic neurosurgical technique, which plays an increasingly
important role in the treatment of medically refractory neu-
rological and psychiatric disorders including Parkinson’s
disease (PD), essential tremor (ET) and dystonia [1–4]. The
interventional proceeding relies on the insertion of electro-
des into specific target structures of the brain sending subse-
quent high frequency stimulation impulses by an implanted
brain pacemaker [5]. However, the major challenges of
13
H. U. Kerl et al.
this surgical technique are the precise identification of the
therapeutic target structure and the accurate placement of
the DBS electrodes into the target area of interest, which is
indispensable for the clinical success [2, 6].
The zona incerta nucleus (ZI) is a horizontally elonga-
ted area of grey matter cells in the subthalamic region. The
nucleus separates the lenticular fasciculus from the thalamic
fasciculus (also known as the “field H1 and H2 of Forel”,
[7]). Its cells are very heterogeneous differing widely in
their shape and size. Its chemoarchitecture is also diverse
containing up to 20 different types of neurochemically defi-
ned cells [8]. The ZI is thought to have several physiologi-
cal functions: The rostral part (rZI), attributed with visceral
control [9], is separated from the dorsolaterally located
subthalamic nucleus (STN) by the pallidofugal fibres (PFF,
[10]), crossing the internal capsule (Ci). The caudal part or
the motor component (cZI, [11]) is located posterior to the
STN and extends behind the prelemniscal radiation. The ZI
is located cranially of the substantia nigra [12, 13], dorso-
medially of the internal capsule, and ventrolaterally of the
red nucleus [14, 15]. Due to its location in the midbrain sur-
rounded by different vital structures and its small size, an
adequate visualisation has proven to be difficult [16].
The STN is currently the preferred target structure used
for DBS treatment in patients with advanced PD [17, 18].
Besides, different groups have reported that the stimulation
of the ZI is an alternative target for DBS [19, 20]. Particu-
larly for the cZI a greater therapeutic benefit for suppressing
tremor in patients with PD and ET compared to the stimu-
lation of the STN has been reported [11, 21]. Furthermore,
stimulation of the cZI has shown remarkable benefit in sup-
pressing resting tremor in PD [22], proximal action tremor
in ET [23], and tremor associated with multiple sclerosis
[24].
However, an adequate and reliable visualisation of the
target structure is vital for the accurate placement of the
DBS electrode [2]. Until now, a satisfactory precise identifi-
cation of the target site based on the currently used pre-ste-
reotactic 1.5 Tesla (T) magnetic resonance imaging (MRI)
alone has limitations [25]. Therefore, indirect and ancillary
targeting approaches have been established. These indirect
techniques employ proportional calculation schemata esti-
mating the position of the target structure with the help of
the midcommissural point [26] and other more closely loca-
ted landmarks as the STN or the red nucleus [27–29]. Sup-
plementary MR images superimposed by an atlas-derived
map are used to more precisely define the localisation of the
target area [30]. Additionally, most neurosurgical centres
verify the exact stereotactic position with the help of intra-
operative electrophysiological function-mapping [31].
The direct targeting method uses the stereotactic pre-ope-
rative MRI for the visualisation and the localisation of the
target structure. This direct method facilitates the individu-
Visualisation of the Zona Incerta for Deep Brain Stimulation at 3.0 Tesla
ally optimal electrode placement, considering the anatomi-
cal variability in position, size, and functional segregation
of the target nuclei [32].
Stereotactic targeting data for the subthalamic region
are frequently calculated pre-operatively from heavily T2-
weighted fast spin-echo MRI [13, 33, 34]. With the recent
establishment of 3.0 T MRI and the development of innova-
tive imaging sequences, new diagnostic possibilities for the
visualisation of the target areas emerge [35]. Consequently,
various alternative MR imaging sequences and techniques
have been proposed to improve the visibility of the DBS
target structures, mainly the STN, including quantitative T1
and T2 imaging [36, 37], T2* mapping [38, 39] and suscep-
tibility-weighted imaging (SWI, [40]). However, the most
promising new imaging techniques for the visualisation of
the ZI have not been compared directly.
Therefore, the aim of this study was to evaluate different
promising new MRI methods (sequence and orientation) for
the visualisation of the ZI at 3.0 T. In order to facilitate an
easy implementation of the results for other neuroradiologi-
cal and neurosurgical centres we used commercially avai-
lable sequence as provided and optimised by the vendor
(standard installation protocol).
Material and Methods
Participant Characteristics
Nine healthy volunteers (five male, four female) with a
mean age of 25 years (range 21–28 years) were included in
this study.
Informed consent was obtained from all participants. The
study was approved by the local research ethics committee.
The approval also covers the analysis of other brain regions
of the volunteers not addressed in this study.
Magnetic Resonance Imaging
For MR imaging of all subjects a 3.0 T system (Magnetom
Trio, Siemens Healthcare, Erlangen, Germany) with a 32-
channel receive head coil (Siemens Healthcare, Erlangen,
Germany) was used. The following sequences were acqui-
red: T1-MPRAGE (T1-weighted magnetisation-prepared
rapid gradient-echo), T2-FLAIR (T2-weighted fluid
attenuation inversion recovery), T2*-FLASH2D (T2*-
weighted two-dimensional fast low angle shot magnetic
resonance imaging with a standard bandwidth of 40 kHz),
T2*-FLASH2D-HB (T2*-FLASH with a high bandwidth
of 200 kHz), T2-SPACE (T2-weighted sampling perfection
with application of optimised contrasts using different
flip angle evolutions) and SWI (susceptibility-weighted
imaging). Slice thickness was manually adapted to comple-
57
tely cover the area of interest while maintaining an adequate
acquisition time. Axial slices of the T1-MPRAGE images
were reconstructed in-line with a slice thickness of 1 mm.
Coronal slices of the SWI scan were reconstructed with a
slice thickness of 0.75 mm. Minimum intensity projections
(MIP) of the SWI datasets were reconstructed in-line with a
slice thickness of 9.6 mm. The specific imaging parameters
used are summarised in Table 1.
Qualitative Evaluation of the Data
All viewing and analysis of the acquired datasets was perfor-
med electronically on a DICOM workstation using OsiriX
Imaging Software (http://www.osirix-viewer.com/index.
html).
Two radiologists analysed the delineation of the ZI for
each dataset independently. Raters were allowed to freely
adjust window/level settings, but no automatic preproces-
sing was applied.
First, both readers assessed the delineation of the ZI vs.
the Ci, the ZI vs. the STN, and the ZI vs. the PFF on the
basis of each reader’s own professional judgment using a
6-point scale. The grading was as follows: 5—excellent
delineation; 4—good delineation; 3—moderate delineation;
2—poor delineation; 1—no delineation; 0—no image/ not
evaluable.
Second, the image quality of each sequence was evalua-
ted by consensus of two radiologists in terms of artifacts
using a 6-point scale (5—no artifacts; 4—minimal artifacts;
3—moderate artifacts; 2—significant artifacts; 1—massive
artifacts; 0—no image/not evaluable).
Quantitative Analysis
The acquired sequences were assessed quantitatively using
the Osirix-software.
The mean signal intensity (SI) was measured for all
datasets by manually placing a region of interest (ROI) of
approximately 0.025 cm2 within the ZI and a ROI of appro-
ximately 0.1 cm2 within the Ci considering intra-individual
identical localisation. The average standard deviation of
noise was quantified by manually placing an ROI (appro-
ximately 2.0 cm2) outside the brain and away from phase-
encoding artifacts. Identical ROI sizes were used for all
corresponding images. ROI measurements were repeated
three times and average values were taken. Each dataset
was carefully scrutinised to identify the sections with the
structures of interest.
After obtaining these measurements for each participant
the signal-to-noise ratios (SNR) and the contrast-to-noise
ratios (CNR) for the ZI were calculated for 18 cerebral
hemispheres according to the equations:
13
58 H. U. Kerl et al.

Table 1  Magnetic resonance imaging parameters for each sequence used in this studya
Sequences TR TE TI Flip FOV Matrix Resolution Slice thick- NoA Receiver band- Scan time Slice
(msec) (msec) (msec) angle (°) (mm) (mm) ness (mm) width (kHz) (min) orientation
T1-MPRAGEb 1900   2.26 1900   9 218 × 250 430 × 512 0.49 × 0.49 1 1 199 03:58 Sagittal
T2-FLAIRc tra 9000   95 2500 145 171 × 220 265 × 200 0.43 × 0.43 4 1 201 05:34 Axial
T2*-FLASH2D-   625   30 –   20 189 × 189 384 × 384 0.49 × 0.49 2 1 200 04:00 Axial
HB trad
T2*-FLASH2D trae   625   25 –   30 192 × 192 384 × 384 0.5 × 0.5 2 1   40 04:00 Axial
T2*-FLASH2D cor   625 f
  25 –   30 192 × 192 384 × 384 0.5 × 0.5 2.5 1   40 04:00 Coronal
T2*-FLASH2D sagg   625   25 –   30 192 × 192 384 × 384 0.5 × 0.5 2.5 1   40 04:24 Sagittal
T2-SPACE trah 3200 353 – 120 127 × 229 290 × 384 0.6 × 0.6 0.6 2 434 03:58 Axial
SWI tra i
  28   20 –   15 180 × 240 221 × 320 0.75 × 0.75 1.2 1 120 05:04 Axial
SWI-MIP traj   28   20 –   15 180 × 240 221 × 320 0.75 × 0.75 9.6 1 120 05:04 Axial
a
TR  time of repetition, TE time of echo, TI inversion time, FOV field of view, NoA numbers of averages, 2D 2 dimensional
b
T1-MPRAGE T1-weighted magnetisation-prepared rapid gradient-echo
c
FLAIR tra  transversal T2-weighted fluid attenuation inversion recovery
d
T2*-FLASH2D-HB tra  transversal T2*-FLASH with a high bandwidth of 200 kHz
T2*-FLASH2D tra  transversal T2*-weighted two-dimensional fast low angle shot magnetic resonance imaging with a standard bandwidth of
e

40 kHz
T2*-FLASH2D cor  coronal T2*-weighted two-dimensional fast low angle shot magnetic resonance imaging with a standard bandwidth of
f

40 kHz
g
T2*-FLASH2D sag  sagittal T2*-weighted two-dimensional fast low angle shot magnetic resonance imaging with a standard bandwidth of
40 kHz
h
T2-SPACE tra  transveral T2-weighted sampling perfection with application of optimised contrasts using different flip angle evolutions
i
SWI tra  transversal susceptibility-weighted imaging
j
SWI-MIP tra  transversal of susceptibility-weighted imaging in minimum intensity projection

SNR = SI ZI /σ , A p-value of 0.05 was used to indicate a statistical

CNR = (SI ZI − SI Ci )/σ ,
where SIZI represents the measured signal (mean) within the
grey matter target structure (ZI), SICi the MRI signal value
in the white matter tracts (internal capsule), and σ the ave-
rage standard deviation of the noise.
To balance the differences of the vendor-optimised
sequences in slice thickness and pixel size the measu-
red SNR and CNR values were adjusted to a voxel of
1 mm × 1 mm × 1 mm.
Statistical Methods
Statistical calculations were performed using the Statistical
Package for the Social Sciences software (SPSS 19, IBM
Corporation, Somers, NY, USA). The inter-rater reliability
for the delineation of the ZI in the healthy volunteer group
was assessed using Cohen’s kappa coefficient (κ), [41].
Kappa values were classified as < 0 indicating no agree-
ment, 0–0.20 as slight, 0.21–0.40 as fair, 0.41–0.60 as
moderate, 0.61–0.80 as substantial, and 0.81–1 as almost
perfect agreement [42, 43].
significance.
Differences in delineation, SNR, and CNR were statisti-
cally evaluated using a paired t-test.
Co-Registration of the Coronal T2*-FLASH2D Images
with the Coronal Section Schema of the Schaltenbrand
and Wahren Atlas
The coronal T2*-FLASH2D image 2 mm anterior to the
midcommissural point was fused with the corresponding
digitised coronal section of the Schaltenbrand and Wahren
stereotactic atlas (plate 27), [14] at the level of the ZI. For
the co-registration clearly delineated anatomic structures
(i.e. SN, globus pallidus, and the wall of the lateral and third
ventricle) in the MR image were used.
Results
Qualitative Results
Exemplary gradings of the image quality in terms of arti-
facts are provided in Fig. 1.

13
Visualisation of the Zona Incerta for Deep Brain Stimulation at 3.0 Tesla 59

Fig. 1 Exemplary illustration of the utilised 6-point scale of the image quality in terms of artifacts (a grade 5 no artifacts (T1-MPRAGE); b grade
4 minimal artifacts (T2-FLAIR), c grade 3 moderate artifacts (SWI))

All acquired sequences provided a good to excellent Table 2 Average artifacts at the level of zona incerta nucleus for each
image quality and none of the images had to be excluded sequencea
due to disturbing artifacts. No artifacts within the region of Sequence Mean ± SD
interest were visible in T1-MPRAGE, minimal artifacts in (consensus of two readers)b
T2-FLAIR, and T2*-FLASH2D imaging, while moderate T1-MPRAGE 5.00 ± 0.00
artifacts were seen in the T2*-FLASH2D-HB, T2-SPACE, T2-FLAIR tra 4.33 ± 0.50
and SWI (SWI-MIP and SWI, Table 2). T2*-FLASH2D-HB tra 3.00 ± 0.00
The first survey of the dataset showed that only the rost- T2*-FLASH2D tra 4.00 ± 0.00
ral part of the ZI could be identified whereas the caudal part T2*-FLASH2D cor 4.00 ± 0.00
was not visible in any of the sequences. For the subsequent T2*-FLASH2D sag 3.78 ± 0.67
analysis we therefore concentrated on the distinguishable T2-SPACE tra 3.00 ± 0.00
rostral part. SWI tra 3.33 ± 0.50
Axial images of a representative healthy volunteer at the SWI cor 2.89 ± 0.33
level of the ZI are shown in Fig. 2. Qualitative ratings are
SWI-MIP tra 3.33 ± 0.50
summarised in Fig. 3.
The T1-weighted images had besides an excellent con- a
SD  standard deviation
trast between cortex and white matter only little contrast b
Artifacts at the level of the zona incerta nucleus for each sequence
within the midbrain regions, especially the ZI. The T2- were evaluated by consensus of two radiologists using a 6-scale
rating system (5—no artifacts; 4—minimal artifacts; 3—moderate
weighted images demonstrated good contrast for the mid- artifacts; 2—significant artifacts; 1—massive artifacts; 0—no
brain structures. Nevertheless, no clear delineation between image/not evaluable)
the ZI and the neighboring structures, the Ci, the STN and
the PFF was seen in the axial T2-FLAIR, T2*-FLASH2D- T2-FLAIR images using a paired t-test. Moreover, the axial
HB, T2-SPACE, and SWI. Axial T2*-FLASH2D provided T2*-FLASH2D sequences showed a slightly superior, sta-
a slightly better demarcation of the ZI vs. the STN. In con- tistically significant delineation of the rZI vs. the STN com-
trast, the T2*-FLASH2D images in coronal orientation, pared to the T2-FLAIR acquisition.
and to a lesser degree in sagittal orientation showed a clear The rZI was fully discernible in coronal T2*-FLASH2D
boundary dividing the rZI from the Ci, the STN, and the images in all nine healthy volunteers in both hemispheres.
PFF (Fig. 4). Similar results with a good delineation of the Exemplary gradings of the delineation are provided in
rZI vs. the adjacent structures could be seen for coronal Figs. 2, 4 and 5.
SWI images (Fig. 5). Furthermore, the coronal and sagit- To analyze possible effects of bandwidth variation on
tal T2*-FLASH2D as well as the coronal SWI sequence the delineation of the ZI we compared two T2*-FLASH2D
provided also statistically a significant superior delineation sequences with bandwidths of 40 kHz (standard installa-
of the rZI vs. the Ci, the STN, and the PFF compared to tion) and 200 kHz (HB). We found that the T2*-FLASH2D

13
60
Fig. 2 Representative axial images of a healthy volunteer at the level
of the zona incerta (top) and enlarged representation of the mesence-
phalon at the level of the zona incerta (bottom) for all sequences (a T1-
MPRAGE; b T2-FLAIR; c T2*-FLASH2D-HB; d T2*-FLASH2D;
e T2-SPACE; f SWI; g SWI-MIP). In this individual the delineation
with 40 kHz provided a clearly superior delineation of the
STN compared to the high bandwidth of 200 kHz.
Concerning inter-rater reliability, both readers graded
the delineation of the ZI vs. the Ci, the STN, and the PFF
identical in T1-MPRAGE, T2-FLAIR, and SWI-MIP ima-
ging with entire agreement. The inter-rater reliability was
at least moderate for the (κ = 1) coronal T2*-FLASH2D
and axial T2-SPACE images. At least fair agreement was
seen for axial and sagittal T2*-FLASH2D, and axial SWI
images. The evaluation of the inter-rater reliability of the
13
H. U. Kerl et al.
of the zona incerta versus the internal capsule, the subthalamic nuc-
leus, and the pallidofugal fibres was graded 1/1/1 for T1-MPRAGE;
2/1/1 for T2-FLAIR; 2/1/1 for T2*-FLASH2D-HB; 2/2/1 for T2*-
FLASH2D; 2/1/1 for T2-SPACE; 2/1/1 for SWI; 2/1/1 for SWI-MIP
T2*FLASH2D-HB and coronal SWI images showed vary-
ing results between slight and perfect agreement depending
on the structure to be delimited (Fig. 3).
Quantitative Results
To provide directly clinical relevant information, we analy-
sed the SNR and CNR of the sequences as provided by the
manufacturer (non-adjusted). Additionally, we adjusted the
Visualisation of the Zona Incerta for Deep Brain Stimulation at 3.0 Tesla 61

Delineation ZI vs. Ci 5 (*)

Inter-rater reliability
Sequence Mean ± SD (Cohen’s kappa {K, p- 4 (*)

Delineation ZI vs. Ci
value}) (*)
3
T1-MPRAGE 1.00 ± 0.00 1.00 (p<0.001)
T2-FLAIR tra 2.39 ± 0.49 0.77 (p=0.001)
2
T2*-FLASH2D-HB tra 2.28 ± 0.45 0.17 (p=0.473)
T2*-FLASH2D tra 2.39 ± 0.49 0.30 (p=0.205)
1
T2*-FLASH2D cor 4.61 ± 0.49 0.54 (p=0.019)
T2*-FLASH2D sag 3.14 ± 0.35 0.31 (p=0.180)
0
T2-SPACE tra 2.00 ± 0.00 1.00 (p<0.001)
SWI tra 2.08 ± 0.28 0.64 (p=0.004)

T1

T2 RA

T2

T2 A S

T2 A S

T2 A S

T2 ASH cor

SW ACE sag

SW a

SW r
-M

-F

*-

*-

*-

*-

-S
SWI cor 3.50 ± 0.51 0.33 (p=0.157)

It

Ic

I-M
FL ra

FL 2D

FL 2D tra

FL 2D
LA E

P
P

IP
IR
SWI-MIP tra 2.00 ± 0.00 1.00 (p<0.001)

tra
G

H
t

tra
2D
a

-H

tra
B
Delineation ZI vs. STN 5 (*)

Inter-rater reliability
(Cohen’s kappa) (K, p- 4
Delineation ZI vs. STN
Sequence Mean ± SD
value)) (*)
3 (*)
T1-MPRAGE 1.00 ± 0.00 1.00 (p<0.001)
T2-FLAIR tra 1.00 ± 0.00 1.00 (p<0.001)
2 (*)
T2*-FLASH2D-HB tra 1.22 ± 0.42 0.36 (p=0.130)
T2*-FLASH2D tra 1.56 ± 0.50 0.78 (p=0.001)
1
T2*-FLASH2D cor 4.53 ± 0.51 0.68 (p=0.002)
T2*-FLASH2D sag 2.58 ± 0.50 0.47 (p=0.019)
0
T2-SPACE tra 1.17 ± 0.38 0.46 (p=0.021)
SWI tra 1.17 ± 0.38 0.27 (p=0.097)
T1

T2

T2

T2

T2

T2

T2

SW

SW a

SW r
-M

-F

*-

*-

*-

*-

-S
SWI cor 3.08 ± 0.28 0.64 (p=0.004)

It

Ic

I-M
FL

FL

FL

FL
LA

PA
PR

IP
AS

AS

AS

AS
IR

C
AG

SWI-MIP tra 1.00 ± 0.00 1.00 (p<0.001)

tra
H

H
tra

tra
2D

2D

2D

2D
E

b
-H

tra

co

sa
B

g
tra

Delineation ZI vs. PFF 5

(*)
Inter-rater reliability
(Cohen’s kappa) (K, p- 4
Delineation ZI vs. PFF

Sequence Mean ± SD
value))
3 (*)
T1-MPRAGE 1.00 ± 0.00 1.00 (p<0.001) (*)
T2-FLAIR tra 1.00 ± 0.00 1.00 (p<0.001)
2
T2*-FLASH2D-HB tra 1.00 ± 0.00 1.00 (p<0.001)
T2*-FLASH2D tra 1.00 ± 0.00 1.00 (p<0.001)
1
T2*-FLASH2D cor 4.19 ± 0.40 0.49 (p=0.016)
T2*-FLASH2D sag 2.22 ± 0.42 0.40 (p=0.034)
0
T2-SPACE tra 1.00 ± 0.00 1.00 (p<0.001)
SWI tra 1.08 ± 0.28 0.64 (p=0.004)
T1

T2

T2

T2 AS

T2 AS

T2 AS

T2 ASH cor

SW ACE sag

SW a

SW r
-M

-F

*-

*-

*-

*-

-S

SWI cor 2.33 ± 0.48 0.05 (p=0.800)

It

Ic

I-M
FL ra

FL 2D

FL 2D tra

FL 2D
LA E

P
PR

IP
IR
AG

SWI-MIP tra 1.00 ± 0.00 1.00 (p<0.001)

co
H

H
t

tra
2D

c
-H

tra
B

Fig. 3 Average delineation of the rostral part of the zona incerta (ZI)
vs. the internal capsule (Ci) (a), vs. the subthalamic nucleus (STN)
(b), and vs. pallidofugal fibres (PFF) (c) for all sequences. The table
(left) indicates the average delineation and the standard deviation of
the mean for each sequence as well as the inter-rater reliability (κ)
with its statistical significance (p-value) for each sequence. The dia-
gram (right) demonstrates the average delineation for the ZI, error bars
indicate the 95%-confidence interval of the mean. Sequences with a
statistically significant superior delineation compared to T2-FLAIR
imaging (paired t-test) are denoted (*)

13
62
Fig. 4 Axial (a), coronal (b), and sagittal (c) views of the subthalamic
region (top) with magnified representation of the area of interest (bot-
tom) in T2*-FLASH2D sequences of a healthy volunteer. The zona
incerta and the surrounding structures are indicated (CCS: splenium
corporis callosi, GP: globus pallidus; PFF: pallidofugal fibres, RN:
Fig. 5  Magnified coronal views of the subthalamic region for T2*-
FLASH2D (a) and SWI (b) of a healthy volunteer. The rostral part
of the zona incerta and its neighboring structures are indicated (GP:
globus pallidus; PFF: pallidofugal fibres, SN: substantia nigra, STN:
measurements for a comparison independent of the acquired
voxel volume.
SNR and CNR measurements for the rZI are shown in
Table 3.
In the non-adjusted measurements the SNR of the rZI in
SWI-MIP, SWI, and T1-MPRAGE images was substanti-
ally higher than in T2*-FLASH2D, T2*-FLASH2D-HB,
T2-FLAIR, and T2-SPACE images and statistically higher
compared to T2-FLAIR imaging. The adjusted SNR values
13
H. U. Kerl et al.
red nucleus, SN: substantia nigra, STN: subthalamic nucleus; ZI: zona
incerta (rostral part)). In this individual the delineation of the zona
incerta versus the internal capsule, the subthalamic nucleus, and the
pallidofugal fibres was graded 2/1/1 for the axial, 5/4/4 for the coronal,
and 3/3/2 for the sagittal orientation
subthalamic nucleus; ZI: zona incerta (rostral part)). In this individual
the delineation of the zona incerta versus the internal capsule, the sub-
thalamic nucleus, and the pallidofugal fibres was graded 5/4/4 for the
coronal T2*-FLASH2D, 4/3/2 for the coronal SWI
differ slightly: Similarly, high SNR values are computed for
T1-MPRAGE and SWI imaging. In addition, T2-SPACE
provided an observable higher SNR compared to T2-FLAIR,
T2*-FLASH2D and T2*-FLASH2D-HB imaging. The
lowest SNR was seen for sagittal T2*-FLASH2D imaging.
After the SNR calculation all scans fulfilled the require-
ment of the Rose criterion (SNR values higher than 5 are
needed for the distinction of image features at a 100% cer-
tainty, [44]).
Visualisation of the Zona Incerta for Deep Brain Stimulation at 3.0 Tesla 63

Table 3 Average signal-to-noise ratios (SNR) and contrast-to-noise ratios (CNR) of the zona incerta for each sequence
Sequence SNR CNR
Non-adjusted SNR Adjusted SNR
b
Non-adjusted CNR Adjusted CNR
Mean ± SDa Mean ± SD Mean ± SD Mean ± SD
T1-MPRAGE 230.89 ± 55.92 (*)c 961.63 ± 232.90 (*) 1.08 ± 0.61 4.51 ± 2.54
T2-FLAIR tra 64.47 ± 12.24 87.16 ± 16.55 4.99 ± 1.53 6.75 ± 2.07
T2*-FLASH2D-HB tra 52.50 ± 7.30 109.33 ± 15.20 (*) 5.62 ± 2.93 11.72 ± 6.10 (*)
T2*-FLASH2D tra 51.89 ± 9.67 103.79 ± 19.33 (*) 6.55 ± 1.67 13.09 ± 3.34 (*)
T2*-FLASH2D cor 56.22 ± 15.88 89.95 ± 20.41 7.86 ± 1.07 (*) 12.57 ± 1.71 (*)
T2*-FLASH2D sag 52.04 ± 9.78 83.27 ± 15.65 6.69 ± 1.87 10.70 ± 2.99 (*)
T2-SPACE tra 53.83 ± 13.36 249.19 ± 61.85 (*) 6.53 ± 1.61 (*) 30.23 ± 7.45 (*)
SWI tra 135.93 ± 39.09 (*) 201.38 ± 57.91 (*) 12.54 ± 7.78 (*) 18.58 ± 11.53 (*)
SWI cor 134.17 ± 33.89 (*) 198.77 ± 50.21 (*) 12.23 ± 8.18 (*) 18.12 ± 12.11 (*)
SWI-MIP tra 499.23 ± 90.77 (*) 92.45 ± 16.81 52.00 ± 17.27 (*) 9.63 ± 3.20
a
Mean and standard deviation (SD)
b
To balance the differences of the vendor optimised sequences in slice thickness and pixel size the measured SNR and CNR values were
adjusted to a voxel of 1 x 1 x 1 mm³
c
Sequences with a statistically significant superior SNR and CNR values compared to T2-FLAIR imaging (paired t-test) are denoted (*)

The lowest non-adjusted CNR values were calculated Discussion

for T1-MPRAGE images, followed by T2-FLAIR, T2*-
FLASH2D-HB, and T2-SPACE images with only minor
variation between the obtained CNR values. Slightly higher
directly measured CNR results were seen for T2*FLASH2D
images. The best non-adjusted CNR values were present in
SWI-MIP and SWI imaging, significantly higher compared
to T2-FLAIR imaging.
After converting the calculated values to an adjusted
voxel volume of 1 mm3, the highest CNR values were com-
puted for T2-SPACE and axial as well as coronal SWI ima-
ges with significant higher values compared to T2-FLAIR
imaging. Also T2*-FLASH2D and T2*-FLASH2D-HB
imaging provided significant higher adjusted CNR values
compared to T2-FLAIR imaging. T1-MPRAGE provided
the lowest adjusted CNR values of all sequences.
Co-Registration of the Coronal Section of the
Schaltenbrand and Wahren Atlas with the Coronal
T2*-FLASH2D-Imaging
Figure 6 shows an example of the co-registration of a coro-
nal T2*-FLASH2D image at the level of the ZI, 2 mm ante-
rior to the midcommissural point, with the most widely used
atlas for brain surgery, the Schaltenbrand and Wahren atlas
for stereotaxy of the human brain [14]. With the help of the
schema the ZI is easily identified. The nucleus clearly pro-
jects in the borders of the ZI demonstrated in the stereotac-
tical schema.
DBS is a reversible stereotactic long-term therapeutic option
for patients with chronic movement disorders including PD
[1, 45], and has been approved by the FDA (US Food and
Drug Administration). For this technique the precise and
reliable localisation of the target structure is indispensable
[2, 6].
In the era of 1.5 T imaging the stereotactic planning was
predominantly done with indirect targeting techniques, as
the SNR and with that the delineation was insufficient for
a direct approach [25, 46]. Recent improvements in MRI
technology, particularly the clinical implementation of hig-
her magnetic field strength (3.0 T) with advances in signal,
contrast, and resolution [47] have increasingly influenced
the stereotactic procedures. This technical advance may
facilitate the inter-individual anatomic variability in posi-
tion, functional segregation, and size of the target struc-
ture [25]. Hence, lately published reports of pre-operative
3.0 MR imaging in PD patients undergoing DBS surgery
have presented encouraging results for the direct targeting
of the STN [48, 49].
However, the recent studies have documented an excel-
lent outcome of PD and ET patients after DBS surgery of a
new target—the ZI [19, 20]. Particularly the stimulation of
the caudal component of the ZI [50, 51] seems to provide
a greater therapeutic benefit compared to the stimulation
of the STN, the currently preferred target structure for PD
[21]. Several studies suggested that cZI nucleus is effective
in suppressing tremors of various etiologies of both the dis-
tal and proximal extremities [52–54]. A recent study com-
paring DBS of the STN and the ZI found similar results for
the motor function but specific improvement in anxiety and

13
64
Zona incerta
Pallidofugal fibres
Subthalamic nucleus
Substantia nigra
a b
Fig. 6 Co-registration of the Schaltenbrand and Wahren atlas for ste-
reotaxy of the human brain with the coronal T2*-FLASH2D-acquisi-
tion of a healthy volunteer at the level of the zona incerta (a). Digitised
coronal schema of the Schaltenbrand and Wahren atlas, depicting a
level 2 mm anterior to the midcommissural point (Fa 2.0) with the in-
self-reported depression for ZI stimulation [55]. However,
the promising new target, the ZI, which is located in the
subthalamic region of the midbrain adjacent to a number
of vital structures [12, 13, 16], is difficult to visualize and
localize.
Therefore, the aim of our study was to determine the
currently optimal sequence and orientation for visualisation
and targeting of the ZI using a standard installation protocol
at 3.0 T.
An important finding of our analysis is that only the ros-
tral part of the ZI could be identified. This has to be consi-
dered when performing DBS planning on the basis of the
sequences used (the rZi may be used as a landmark for indi-
rect targeting of the cZi).
We suppose that the discrepant identifiability is secon-
dary to differences in histological composition that also
become obvious in the varying appearance on the histologi-
cal sections of the stereotactic atlas.
Our qualitative results show that the rZI was best demar-
cated on T2*-FLASH2D and SWI imaging. Even though the
axial T2-weighted MRI (T2-FLAIR, T2*-FLASH2D-HB,
and T2-SPACE) provided a good visualisation of midbrain
region, the delineation of the rZI and the adjacent structu-
res was limited. T1-MPRAGE demonstrated only minimal
distinction of the midbrain structures with no discernible
delineation of the rZI.
Concerning the orientation, the coronal view of the T2*-
FLASH2D and SWI images and to a lesser degree the sagit-
tal view of the T2*-FLASH2D acquisition provided the
optimal delineation of the rZI vs. the Ci, the STN, and the
PFF. In contrast, neither axial SWI, nor axial T2*-FLASH2D
imaging facilitated a delineation of the rZI in all cases.
13
H. U. Kerl et al.
Thalamic nuclei
Internal capsule Putamen
Globus pallidus externus
Globus pallidus internus
dication of the main cerebral structures in the area of interest (b). The
zona incerta (ZI) is clearly visible as a small biconvex, hypointense
structure within the boundaries of the Schaltenbrand and Wahren atlas
schema for the ZI
A major finding of the present study is that the rZI was
optimally visualised on T2*-FLASH2D and SWI imaging
in coronal orientation. This may for the first time facilitate a
visualisation of the rZI as a basis for MRI-guided targeting
approaches.
A sufficient inter-rater reliability of neuroimaging is cru-
cial for neurosurgical procedures. We determined the inter-
rater reliability of two readers for the delineation of the rZI
vs. Ci, the STN, and vs. the PFF using Cohen’s kappa. The
inter-rater reliability of the most adequate sequence for the
rZI was at least moderate for the coronal T2*-FLASH2D
and varied from slight to substantial for the coronal SWI
depending on the structure to be delimited.
These values indicate that using appropriate sequences at
3.0 T a reliable delineation of the rZI performed by readers
experienced in the imaging of the deep brain nuclei is pos-
sible. However, further optimisation of the sequences has to
be accomplished to ameliorate the inter-rater agreement.
To quantify the intrinsic quality of the acquired images
the SNR and CNR measurements for the rZI were conduc-
ted. In general, SNR and CNR values are linear proportio-
nal to slice thickness and pixel size [56]. Due to the use of
standard parameters with an acquisition time appropriate for
clinical conditions, the slice thickness and pixel size differ
between the sequences used. To additionally obtain geome-
trically comparable values the SNR and CNR measurements
were adjusted to a voxel of 1  × 1 × 1 mm3. The geometric
correction has been used previously to compare vendor-
optimised sequences with different voxel sizes [57].
Our quantitative analysis provided the highest non-adjus-
ted CNR values for the susceptibility-weighted sequences
(SWI-MIP and SWI), while T1-MPRAGE show by far the
lowest CNR results for the rZI. The highest adjusted CNR
Visualisation of the Zona Incerta for Deep Brain Stimulation at 3.0 Tesla
Fig. 7 The use of susceptibility-weighted imaging for stereotactic
neurosurgery planning. The deep cerebral veins (ICV: internal cerebral
veins, TSV: thalamostriate veins, SV: septal veins) can be easily identi-
results were calculated for SWI and T2-SPACE imaging.
For T2*-FLASH2D sequences significant higher adjusted
CNR values compared to T2-FLAIR imaging were obtai-
ned. The slight CNR changes between non-adjusted and
adjusted values can be attributed to several factors:
Firstly, the CNR adjustments for the high resolution
sequence T2-SPACE results in a high converting factor and
consecutively in high CNR values.
Secondly, the SD of noise in the T2*-FLASH2D sequen-
ces was relatively high compared to the other sequences.
This sequence characteristic also contributes to a low CNR
values for T2*-FLASH2D in comparison to T2-SPACE.
Thirdly, CNR changes due to intra-voxel signal decrease
are neglected in the geometric voxel adjustment [58].
Summing up, both the qualitative and the quantitative
analysis of the rZI show that sequences that are susceptible
to local field inhomogeneities provide a good delineation of
the rZI. The chemical composition of the ZI is not exactly
known. The signal characteristics of adjacent nuclei as the
STN are probably caused by iron deposition [13].
In general, the involvement of iron in many neurodegene-
rative diseases is well documented [59, 60] and an increase
in iron concentration of the basal ganglia with age is known
[61]. Particularly, in deep brain structures of PD patients a
progressive increase of iron concentration has been repor-
ted [62, 63]. Although, the exact correlation between iron
accumulation and neurodegenerative diseases still remains
unclear, recent studies suggested that this finding is a result
of a non-specific effect of neuronal degeneration [13].
The high iron concentration corresponds to a hypointense
signal of the STN and internal globus pallidus in gradient
echo sequences with T2*-weighted images and susceptibi-
65
fied at different brain levels (a–c) so that the trajectory can be adjusted
to avoid haemorrhage
lity weighted images [13]. SWI provides a combination of
gradient-echo magnitude and phase change images influen-
ced by the magnetic susceptibility of iron [64, 65]. We the-
refore assume that the signal characteristics of the ZI may
analogously be caused by iron deposition.
The supposed iron deposition in the ZI may render pati-
ents affected by neurodegenerative diseases especially sui-
table for T2*, T2*-FLASH2D and SWI imaging.
As an additional benefit, SWI provided a clear visualisa-
tion of deep cerebral veins and transparenchymal vessels.
These anatomic details are of special interest for the pre-
operative planning of the DBS surgery. Hence, predictable
intracranial haemorrhages due to blood vessel injuries can
be avoided (Fig. 7).
Based on the rationale that a higher bandwidth should
influence the appearance of iron-containing structures [66]
due to a reduction of susceptibility artifacts [67], we perfor-
med an initial comparison of two T2*-FLASH2D sequen-
ces with different bandwidths for the visualisation of the
ZI. Whereas there was no major difference in quantitative
values between the two sequences, in the qualitative analy-
sis we did observe a better delineation of the rZI using the
low-bandwidth T2*-FLASH2D sequence. This finding is
consistent with prior studies indicating more susceptibility
artifacts at lower bandwidths and therefore a better visuali-
sation of iron-rich structures [68, 69]. For the imaging of
the ZI in patients affected by neurodegenerative disorders
however, fewer artifacts might be desirable as higher iron
content can be assumed.
Future studies are requisite to determine the optimal
bandwidth for “FLASH” sequences in ZI imaging.
13
66
Our study has a few potential limitations. First, we stu-
died a population of young and healthy adults. This impli-
cates that slightly different results have to be expected in
patients with neurodegenerative diseases (see above).
Second, the comparison of different vendor-optimised
sequences entailed the comparison of different voxel sizes.
We additionally adjusted the voxel size to study the intrinsic
sequence characteristics. This geometric adjustment ignores
the intra-voxel signal decay (see above). However, the pri-
mary use of sequences adjusted to similar voxel size would
have resulted in long scanning times and impractical slice
thicknesses, which was not within the intent of our study.
Although the rZI is well visualised on T2*-FLASH2D
and susceptibility-weighted imaging and the CNR and SNR
are high for the region of interest, the anatomic accuracy of
the approach had to be confirmed. In a co-registration of the
left hemisphere of the coronal T2*-FLASH2D image with
the Schaltenbrand and Wahren atlas the ZI obviously pro-
jects in the borders of the stereotactical schema. This further
confirms the correct identification and localisation of the ZI
and its surrounding structures.
However, despite the superior delineation of the rZI in
T2*-FLASH2D and SWI images, the results cannot be
transferred directly to clinical practice. Confirmation of our
findings in a patient group remains to be performed. Verifi-
cation of the geometric accuracy for the direct targeting of
the ZI based on 3.0 T MR imaging alone should be under-
taken in advance to clinical implementation. Additionally,
validation of the imaging results with post-mortem analysis
should be performed to confirm histologically that the struc-
ture referred to as ZI/rZI truly corresponds to the nucleus
and its full extent.
Furthermore, motion artifacts have to be expected in a
PD and ET patient’s group (if not performed under general
anaesthesia). The artifacts will influence the delineation of
the rZI and limit the image quality.
Conclusion
In the current study, we have shown that using a standard
installation protocol at 3.0 T the caudal zone of the ZI could
not be delineated, the visualisation of the rostral zone was
limited but possible using the T2*-FLASH2D, particularly
in the coronal view.
Furthermore, susceptibility-weighted imaging clearly
visualizes the deep cerebral veins and transparenchymal
vessels and is therefore of additional importance for pre-
operative DBS planning. In addition, through the use of
vendor-optimised sequences our protocol can be easily
implemented in clinical routine.
In summary, for clinical practice a combination of both
techniques, T2*-FLASH2D and SWI, may lead to a more
13
H. U. Kerl et al.
precise ZI targeting, shorten the intervention time and pos-
sibly decrease the frequency of complicating haemorrhages
in DBS surgery.
Conflict of Interest  The authors declare that there is no actual or
potential conflict of interest in relation to this article.
References
  1. The deep-brain stimulation for Parkinson’s disease study group.
Deep-brain stimulation of the subthalamic nucleus or the pars
interna of the globus pallidus in Parkinson’s disease. N Engl J
Med. 2001 Sep 27;345(13):956–63.
  2. Plaha P, Javed S, Agombar D, Genevive OF, Khan S, Whone A, et
al. Bilateral caudal zona incerta nucleus stimulation for essential
tremor: outcome and quality of life. J Neurol Neurosurg Psychia-
try. 2011 Feb 1;82:899–904.
  3. Tisch S, Rothwell JC, Limousin P, Hariz MI, Corcos DM. The
physiological effects of pallidal deep brain stimulation in dysto-
nia. IEEE Trans Neural Syst Rehabil Eng. 2007;15(2):166–72.
  4. Deuschl G, Schade-Brittinger C, Krack P, Volkmann J, Schafer H,
Botzel K, et al. A randomised trial of deep-brain stimulation for Par-
kinson’s disease. N Engl J Med. 2006 Aug 31;355(9):896–908.
  5. Kringelbach ML, Jenkinson N, Owen SL, Aziz TZ. Transla-
tional principles of deep brain stimulation. Nat Rev Neurosci.
2007;8(8):623–35.
  6. McIntyre CC, Savasta M, Kerkerian-Le Goff L, Vitek JL.
Uncovering the mechanism(s) of action of deep brain sti-
mulation: activation, inhibition, or both. Clin Neurophysiol.
2004;115(6):1239–48.
  7. Mitrofanis J. Some certainty for the “zone of uncertainty”?
Exploring the function of the zona incerta. Neuroscience.
2005;130(1):1–15.
  8. Mitrofanis J, Ashkan K, Wallace BA, Benabid AL. Chemoarchi-
tectonic heterogeneities in the primate zona incerta: clinical and
functional implications. J Neurocytol. 2004;33(4):429–40.
  9. Mok D, Mogenson GJ. Convergence of signals in the zona incerta
for angiotensin-mediated and osmotic thirst. Brain Res. 1987 Mar
31;407(2):332–40.
10. Parent M, Parent A. The pallidofugal motor fiber system in prima-
tes. Parkinsonism Relat Disord. 2004;10(4):203–11.
11. Plaha P, Khan S, Gill SS. Bilateral stimulation of the caudal zona
incerta nucleus for tremor control. J Neurol Neurosurg Psychia-
try. 2008;79(5):504–13.
12. Carpenter M. The subthalamic region. In: Carpenter M, editor.
Human Neuroanatomy. Baltimore: Wiliams & Wilkins; 1976.
pp. 509–11.
13. Dormont D, Ricciardi KG, Tande D, Parain K, Menuel C, Galanaud
D, et al. Is the subthalamic nucleus hypointense on T2-weighted
images? A correlation study using MR imaging and stereotactic
atlas data. AJNR Am J Neuroradiol. 2004;25(9):1516–23.
14. Schaltenbrand G, Wahren W. Atlas for stereotaxy of the human
brain. 2nd ed. Stuttgart: Thieme; 1977.
15. Yelnik J, Percheron G. Subthalamic neurons in prima-
tes: a quantitative and comparative analysis. Neuroscience.
1979;4(11):1717–43.
16. Hattingen E, Blasel S, Nichtweiss M, Zanella FE, Weidauer
S. MR imaging of midbrain pathologies. Clin Neuroradiol.
2010;20(2):81–97.
17. Castrioto A, Lozano AM, Poon YY, Lang AE, Fallis M, Moro E.
Ten-year outcome of subthalamic stimulation in Parkinson disease:
a blinded evaluation. Arch Neurol. 2011 Aug 8;68(12):1550–6.
Visualisation of the Zona Incerta for Deep Brain Stimulation at 3.0 Tesla
18. Mandat T, Tykocki T, Koziara H, Koziorowski D, Brodacki B,
Rola R, et al. Subthalamic deep brain stimulation for the treatment
of Parkinson disease. Neurol Neurochir Pol. 2011;45(1):32–6.
19. Lanotte MM, Rizzone M, Bergamasco B, Faccani G, Melcarne
A, Lopiano L. Deep brain stimulation of the subthalamic nuc-
leus: anatomical, neurophysiological, and outcome correlations
with the effects of stimulation. J Neurol Neurosurg Psychiatry.
2002;72(1):53–8.
20. Voges J, Volkmann J, Allert N, Lehrke R, Koulousakis A, Freund
HJ, et al. Bilateral high-frequency stimulation in the subthalamic
nucleus for the treatment of Parkinson disease: correlation of the-
rapeutic effect with anatomical electrode position. J Neurosurg.
2002;96(2):269–79.
21. Plaha P, Ben-Shlomo Y, Patel NK, Gill SS. Stimulation of the
caudal zona incerta is superior to stimulation of the subthalamic
nucleus in improving contralateral parkinsonism. Brain. 2006;129
(Pt 7):1732–47.
22. Kitagawa M, Murata J, Uesugi H, Kikuchi S, Saito H, Tashiro
K, et al. Two-year follow-up of chronic stimulation of the pos-
terior subthalamic white matter for tremor-dominant Parkinson’s
disease. Neurosurgery. 2005;56(2):281–9; discussion 9.
23. Kitagawa M, Murata J, Kikuchi S, Sawamura Y, Saito H, Sasaki
H, et al. Deep brain stimulation of subthalamic area for severe
proximal tremor. Neurology. 2000 Jul 12;55(1):114–6.
24. Nandi D, Chir M, Liu X, Bain P, Parkin S, Joint C, et al. Elec-
trophysiological confirmation of the zona incerta as a target for
surgical treatment of disabling involuntary arm movements in
multiple sclerosis: use of local field potentials. J Clin Neurosci.
2002;9(1):64–8.
25. Cho ZH, Min HK, Oh SH, Han JY, Park CW, Chi JG, et al. Direct
visualisation of deep brain stimulation targets in Parkinson disease
with the use of 7-tesla magnetic resonance imaging. J Neurosurg.
2010;113(3):639–47.
26. Spiegelmann R, Friedman WA. Rapid determination of thalamic
CT-stereotactic coordinates: a method. Acta Neurochir (Wien).
1991;110(1–2):77–81.
27. Bejjani BP, Dormont D, Pidoux B, Yelnik J, Damier P, Arnulf
I, et al. Bilateral subthalamic stimulation for Parkinson’s
disease by using three-dimensional stereotactic magnetic reso-
nance imaging and electrophysiological guidance. J Neurosurg.
2000;92(4):615–25.
28. Aziz TZ, Nandi D, Parkin S, Liu X, Giladi N, Bain P, et al. Tar-
geting the subthalamic nucleus. Stereotact Funct Neurosurg.
2001;77(1–4):87–90.
29. Murata J, Kitagawa M, Uesugi H, Saito H, Iwasaki Y, Kikuchi
S, et al. Electrical stimulation of the posterior subthalamic area
for the treatment of intractable proximal tremor. J Neurosurg.
2003;99(4):708–15.
30. Andrade-Souza YM, Schwalb JM, Hamani C, Eltahawy H, Hoque
T, Saint-Cyr J, et al. Comparison of three methods of targeting
the subthalamic nucleus for chronic stimulation in Parkinson’s
disease. Neurosurgery. 2005;56(2 Suppl):360–8; discussion -8.
31. Amirnovin R, Williams ZM, Cosgrove GR, Eskandar EN. Expe-
rience with microelectrode guided subthalamic nucleus deep
brain stimulation. Neurosurgery. 2006;58(1 Suppl):ONS96–102;
discussion ONS96-.
32. Ashkan K, Blomstedt P, Zrinzo L, Tisch S, Yousry T, Limou-
sin-Dowsey P, et al. Variability of the subthalamic nucleus:
the case for direct MRI guided targeting. Br J Neurosurg.
2007;21(2):197–200.
33. Kitajima M, Korogi Y, Kakeda S, Moriya J, Ohnari N, Sato T, et
al. Human subthalamic nucleus: evaluation with high-resolution
MR imaging at 3.0 T. Neuroradiology. 2008;50(8):675–81.
67
34. Brunenberg EJ, Platel B, Hofman PA, Ter Haar Romeny BM,
Visser-Vandewalle V. Magnetic resonance imaging techni-
ques for visualisation of the subthalamic nucleus. J Neurosurg.
2011;115(5):971–84.
35. Lanfermann H, Pilatus U, Preibisch C, Hattingen E, Weidauer
S, Bink A, et al. 3 T in der Neuroradiologie—Spannungs-
feld zwischen Wunsch und Wirklichkeit. Clin Neuroradiol.
2006;16(2):91–108.
36. Guo T, Finnis KW, Deoni SC, Parrent AG, Peters TM. Compari-
son of different targeting methods for subthalamic nucleus deep
brain stimulation. Med Image Comput Comput Assist Interv.
2006;9(Pt 1):768–75.
37. Gringel T, Schulz-Schaeffer W, Elolf E, Frolich A, Dechent P,
Helms G. Optimised high-resolution mapping of magnetisation
transfer (MT) at 3 Tesla for direct visualisation of substructures
of the human thalamus in clinically feasible measurement time. J
Magn Reson Imaging. 2009;29(6):1285–92.
38. Duyn JH, van Gelderen P, Li TQ, de Zwart JA, Koretsky AP,
Fukunaga M. High-field MRI of brain cortical substructure
based on signal phase. Proc Natl Acad Sci U S A. 2007 Jul
10;104(28):11796–801.
39. O’Gorman RL, Shmueli K, Ashkan K, Samuel M, Lythgoe DJ,
Shahidiani A, et al. Optimal MRI methods for direct stereotac-
tic targeting of the subthalamic nucleus and globus pallidus. Eur
Radiol. 2011;21(1):130–6.
40. Haacke EM, Xu Y, Cheng YC, Reichenbach JR. Susceptibility
weighted imaging (SWI). Magn Reson Med. 2004;52(3):612–8.
41. Cohen JA. A coefficient of agreement for nominal scales. Educ
Psychol Meas. 1960;20:37–46.
42. Landis JR, Koch GG. An application of hierarchical kappa-type
statistics in the assessment of majority agreement among multiple
observers. Biometrics. 1977;33(2):363–74.
43. Landis JR, Koch GG. The measurement of observer agreement
for categorical data. Biometrics. 1977;33(1):159–74.
44. Bushberg JT, et al. The essential physics of medical imaging. 2nd
ed. Philadelphia: Lippincott Williams & Wilkins; 2006.
45. Ostergaard K, Sunde N, Dupont E. Effects of bilateral sti-
mulation of the subthalamic nucleus in patients with severe
Parkinson’s disease and motor fluctuations. Mov Disord.
2002;17(4):693–700.
46. Cuny E, Guehl D, Burbaud P, Gross C, Dousset V, Rougier A.
Lack of agreement between direct magnetic resonance imaging
and statistical determination of a subthalamic target: the role of
electrophysiological guidance. J Neurosurg. 2002;97(3):591–7.
47. Biswas J, Nelson CB, Runge VM, Wintersperger BJ, Baumann
SS, Jackson CB, et al. Brain tumor enhancement in magnetic
resonance imaging: comparison of signal-to-noise ratio (SNR)
and contrast-to-noise ratio (CNR) at 1.5 versus 3 T. Invest Radiol.
2005;40(12):792–7.
48. Slavin KV, Thulborn KR, Wess C, Nersesyan H. Direct visuali-
sation of the human subthalamic nucleus with 3 T MR imaging.
AJNR Am J Neuroradiol. 2006;27(1):80–4.
49. Toda H, Sawamoto N, Hanakawa T, Saiki H, Matsumoto S, Oku-
mura R, et al. A novel composite targeting method using high-
field magnetic resonance imaging for subthalamic nucleus deep
brain stimulation. J Neurosurg. 2009;111(4):737–45.
50. Hamel W, Fietzek U, Morsnowski A, Schrader B, Herzog J, Wein-
ert D, et al. Deep brain stimulation of the subthalamic nucleus
in Parkinson’s disease: evaluation of active electrode contacts. J
Neurol Neurosurg Psychiatry. 2003;74(8):1036–46.
51. Herzog J, Fietzek U, Hamel W, Morsnowski A, Steigerwald F,
Schrader B, et al. Most effective stimulation site in subthalamic
deep brain stimulation for Parkinson’s disease. Mov Disord.
2004;19(9):1050–4.
13
68
52. Blomstedt P, Sandvik U, Linder J, Fredricks A, Forsgren L, Hariz
MI. Deep brain stimulation of the subthalamic nucleus versus the
zona incerta in the treatment of essential tremor. Acta Neurochir
(Wien). 2011;153(12):2329–35.
53. Fytagoridis A, Sandvik U, Astrom M, Bergenheim T, Blomstedt
P. Long term follow-up of deep brain stimulation of the caudal
zona incerta for essential tremor. J Neurol Neurosurg Psychiatry.
2011 Dec 28.
54. Sandvik U, Hariz GM, Blomstedt P. Quality of life following
DBS in the caudal zona incerta in patients with essential tremor.
Acta Neurochir (Wien). 2011 Nov 23.
55. Burrows AM, Ravin PD, Novak P, Peters ML, Dessureau B,
Swearer J, et al. Limbic and motor function comparison of deep
brain stimulation of the zona Incerta and subthalamic nucleus.
Neurosurgery. 2011 Aug 19.
56. Stark DD, Bradley WG. Magnetic resonance imaging. St. Louis:
C.V. Mosby; 1999.
57. Haneder S, Attenberger UI, Biffar A, Dietrich O, Fink C, Schoen-
berg SO, et al. Gadofosveset: parameter optimisation for steady-
state imaging of the thoracic and abdominal vasculature. Invest
Radiol. 2011 Jun 24.
58. Heise CE, Mitrofanis J. Reduction in parvalbumin expression
in the zona incerta after 6OHDA lesion in rats. J Neurocytol.
2005;34(6):421–34.
59. Kosta P, Argyropoulou MI, Markoula S, Konitsiotis S. MRI eva-
luation of the basal ganglia size and iron content in patients with
Parkinson’s disease. J Neurol. 2006;253(1):26–32.
60. McNeill A, Chinnery PF. Neurodegeneration with brain iron
accumulation. Handb Clin Neurol. 2011;100:161–72.
61. Schicha H, Kasperek K, Feinendegen LE, Siller V, Klein HJ. Iron
content of the human brain and its correlation to age. Beitr Pathol.
1971;142(3):268–74.
13
H. U. Kerl et al.
62. Linn J, Danek A, Hoffmann L, Seelos K, Brückmann H. Diffe-
rential diagnosis of bilateral thalamic lesions. Clin Neuroradiol.
2007;17(1):3–22.
63. Zhang W, Sun SG, Jiang YH, Qiao X, Sun X, Wu Y. Deter-
mination of brain iron content in patients with Parkinson’s
disease using magnetic susceptibility imaging. Neurosci Bull.
2009;25(6):353–60.
64. Haacke EM, Ayaz M, Khan A, Manova ES, Krishnamurthy B, Gol-
lapalli L, et al. Establishing a baseline phase behavior in magnetic
resonance imaging to determine normal vs. abnormal iron content
in the brain. J Magn Reson Imaging. 2007;26(2):256–64.
65. Lummel N, Wiesmann M, Bruckmann H, Linn J. The value of
different magnetic resonance imaging sequences for the detection
of intraventricular hemorrhages. Klin Neuroradiol. 2010 Feb 28.
66. Vertinsky AT, Coenen VA, Lang DJ, Kolind S, Honey CR, Li D,
et al. Localisation of the subthalamic nucleus: optimisation with
susceptibility-weighted phase MR imaging. AJNR Am J Neuro-
radiol. 2009;30(9):1717–24.
67. Abosch A, Yacoub E, Ugurbil K, Harel N. An assessment of current
brain targets for deep brain stimulation surgery with susceptibi-
lity-weighted imaging at 7 T. Neurosurgery. 2010;67(6):1745–56;
discussion 56.
68. Runge VM, Wood ML, Kaufman DM, Traill MR, Nelson KL.
The straight and narrow path to good head and spine MRI. Radio-
graphics. 1988;8(3):507–31.
69. Heiland S. From A as in Aliasing to Z as in Zipper: Artifacts in
MRI. Clin Neuroradiol. 2008;18(1):25–36.