ArticlePDF Available

Accuracy of different three-dimensional subcortical human brain atlases for DBS –lead localisation

Authors:

Abstract and Figures

Background: Accurate interindividual comparability of deep brain stimulation (DBS) lead locations in relation to the surrounding anatomical structures is of eminent importance to define and understand effective stimulation areas. The objective of the current work is to compare the accuracy of the DBS lead localisation relative to the STN in native space with four recently developed three-dimensional subcortical brain atlases in the MNI template space. Accuracy is reviewed by anatomical and volumetric analysis as well as intraoperative electrophysiological data. Methods: Postoperative lead localisations of 10 patients (19 hemispheres) were analysed in each individual patient based on Brainlab software (native space) and after normalization into the MNI space and application of 4 different human brain atlases using Lead-DBS toolbox within Matlab (template space). Each patient's STN was manually segmented and the relation between the reconstructed lead and the STN was compared to the 4 atlas-based STN models by applying the Dice coefficient. The length of intraoperative electrophysiological STN activity along different microelectrode recording tracks was measured and compared to reconstructions in native and template space. Descriptive non-parametric statistical tests were used to calculate differences between the 4 different atlases. Results: The mean STN volume of the study cohort was 153.3 ± 40.3 mm3 (n = 19). This is similar to the STN volume of the DISTAL atlas (166 mm3; p = .22), but significantly larger compared to the other atlases tested in this study. The anatomical overlap of the lead-STN-reconstruction was highest for the DISTAL atlas (0.56 ± 0.18) and lowest for the PD25 atlas (0.34 ± 0.17). A total number of 47 MER trajectories through the STN were analysed. There was a statistically significant discrepancy of the electrophysiogical STN activity compared to the reconstructed STN of all four atlases (p < .0001). Conclusion: Lead reconstruction after normalization into the MNI template space and application of four different atlases led to different results in terms of the DBS lead position relative to the STN. Based on electrophysiological and imaging data, the DISTAL atlas led to the most accurate display of the reconstructed DBS lead relative to the DISTAL-based STN.
Content may be subject to copyright.
Contents lists available at ScienceDirect
NeuroImage: Clinical
journal homepage: www.elsevier.com/locate/ynicl
Accuracy of dierent three-dimensional subcortical human brain atlases for
DBS lead localisation
Andreas Nowacki
a,,1
, T.A-K. Nguyen
a,1
, Gerd Tinkhauser
b,c
, Katrin Petermann
b
, Ines Debove
b
,
Roland Wiest
d
, Claudio Pollo
a
a
Department of Neurosurgery, Inselspital, University Hospital Bern, and University of Bern, Bern, Switzerland
b
Department of Neurology, Inselspital, University Hospital Bern, and University of Bern, Bern, Switzerland
c
Medical Research Council Brain Network Dynamics Unit and Nueld Department of Clinical Neurosciences, University of Oxford, United Kingdom
d
Department of diagnostic and interventional Neuroradiology, Inselspital, University Hospital Bernand University of Bern, Bern, Switzerland
ARTICLE INFO
Keywords:
Deep brain stimulation
Lead localisation
Human brain atlas
MNI space
ABSTRACT
Background: Accurate interindividual comparability of deep brain stimulation (DBS) lead locations in relation to
the surrounding anatomical structures is of eminent importance to dene and understand eective stimulation
areas. The objective of the current work is to compare the accuracy of the DBS lead localisation relative to the
STN in native space with four recently developed three-dimensional subcortical brain atlases in the MNI tem-
plate space. Accuracy is reviewed by anatomical and volumetric analysis as well as intraoperative electro-
physiological data.
Methods: Postoperative lead localisations of 10 patients (19 hemispheres) were analysed in each individual
patient based on Brainlab software (native space) and after normalization into the MNI space and application of
4dierent human brain atlases using Lead-DBS toolbox within Matlab (template space). Each patient's STN was
manually segmented and the relation between the reconstructed lead and the STN was compared to the 4 atlas-
based STN models by applying the Dice coecient. The length of intraoperative electrophysiological STN ac-
tivity along dierent microelectrode recording tracks was measured and compared to reconstructions in native
and template space. Descriptive non-parametric statistical tests were used to calculate dierences between the 4
dierent atlases.
Results: The mean STN volume of the study cohort was 153.3 ± 40.3 mm3 (n= 19). This is similar to the STN
volume of the DISTAL atlas (166 mm3; p= .22), but signicantly larger compared to the other atlases tested in
this study. The anatomical overlap of the lead-STN-reconstruction was highest for the DISTAL atlas
(0.56 ± 0.18) and lowest for the PD25 atlas (0.34 ± 0.17). A total number of 47 MER trajectories through the
STN were analysed. There was a statistically signicant discrepancy of the electrophysiogical STN activity
compared to the reconstructed STN of all four atlases (p< .0001).
Conclusion: Lead reconstruction after normalization into the MNI template space and application of four
dierent atlases led to dierent results in terms of the DBS lead position relative to the STN. Based on elec-
trophysiological and imaging data, the DISTAL atlas led to the most accurate display of the reconstructed DBS
lead relative to the DISTAL-based STN.
1. Introduction
Deep Brain Stimulation (DBS) is a surgical treatment option to al-
leviate symptoms of movement disorders such as Parkinson's Disease
(PD) (Deuschl et al., 2006;Schuepbach et al., 2013). Dierent
subcortical targets within the basal ganglia and thalamus are typically
chosen for DBS lead placement. For instance, the Subthalamic Nucleus
(STN) is a common target for DBS of PD. Evidence suggests accurate
lead placement to be of paramount importance for postoperative out-
come (Welter et al., 2014;Wodarg et al., 2012). Therefore, exact
https://doi.org/10.1016/j.nicl.2018.09.030
Received 6 May 2018; Received in revised form 17 August 2018; Accepted 25 September 2018
Abbreviations: Deep Brain Stimulation, (DBS); microelectrode recording, (MER); Parkinson's Disease, (PD); Subthalamic Nucleus, (STN); MNI152, (Montreal
Neurological Institute 152)
Corresponding author at: Department of Neurosurgery, Inselspital, University Hospital Bern, and University of Bern, Bern, Switzerland.
1
The rst two authors contributed equally to this work.
E-mail address: neuro.nowacki@gmail.com (A. Nowacki).
NeuroImage: Clinical 20 (2018) 868–874
Available online 27 September 2018
2213-1582/ © 2018 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T
localisation of the DBS lead within the target and in relationship to
surrounding anatomical structures may be helpful to analyse the un-
derlying reason for poor stimulation ecacy. Moreover, with advanced
lead design oering directional stimulation and current steering, pro-
gramming can be adjusted according to the DBS lead location and the
intended target structure to minimize stimulation induced side-eects
(Pollo et al., 2014;Schupbach et al., 2017;Tinkhauser et al., 2018).
Apart from DBS lead localisation in an individual patient (native space),
comparison of DBS lead positions across subjects is an important and
powerful tool for scientic approaches, e.g., to dene ecacious sti-
mulation areas (sweet spots) within well-known target structures
(Accolla et al., 2016;Akram et al., 2017).
The undertaking of comparing anatomical data across subjects is
challenging because of anatomical heterogeneity. Both position and
sizes of anatomical structures in one brain must correspond to positions
and sizes in another to make meaningful comparisons (Brett et al.,
2002). With advances in imaging technologies and deformable regis-
tration algorithms, projection of DBS leads onto three-dimensional
subcortical atlases has been established more recently to compare DBS
lead locations of dierent subjects (Cheung et al., 2014;Welter et al.,
2014). The basic underlying workow usually consists of registering
brains of individuals to a common template by applying non-rigid,
deformable registration algorithms (normalization into template space)
(Klein et al., 2009). Currently, the MNI152 (Montreal Neurological
Institute) space, which is based on anatomical high-resolution images of
152 subjects, is broadly used as a template space. Implementation of
brain-atlases into the template space then enables further anatomical
analyses and comparison of interindividual data. Today, many dierent
atlases that have been dened on the basis of magnetic resonance
imaging (MRI) and/or histology are available (Chakravarty et al., 2006;
Welter et al., 2014;Xiao et al., 2015;Yelnik et al., 2007). A major
drawback of many atlases is that their spatial denition of anatomical
structures does not correspond inherently to the MNI template (sum-
marised in Ewert et al. (Ewert et al., 2017)). However, this is a crucial
step for the analysis of lead locations in the context of surrounding
structures. Inaccurate registrations will ultimately lead to wrong con-
clusions about DBS lead placements and the stimulated structures and
can therefore render group comparisons erroneous. Another drawback
is that most atlases are based on young and healthy individuals,
whereas the study groups of interest in the eld of DBS are usually old
and their brains atrophic (Dalaker et al., 2011). To overcome these
limitations, dierent groups have recently presented atlases that match
the MNI space or which are based on PD patients instead of healthy
subjects (Ewert et al., 2017;Pauli et al., 2017;Wang et al., 2016;Xiao
et al., 2015).
The objective of the current work is to compare the accuracy of the
DBS lead location relative to the STN in native space with four recently
developed three-dimensional subcortical brain atlases in the MNI space.
Accuracy is reviewed by anatomical and volumetric analysis as well as
intraoperative electrophysiological data.
2. Material and Methods
2.1. Patients
We retrospectively analysed 24 patients undergoing DBS lead im-
plantation into the STN for advanced PD between October 2015 and
April 2017 at our department. Inclusion criteria were age older than
18 years, availability of intraoperative electrophysiological testing re-
sults by specialized movement disorder neurologists, microelectrode
recording results available along at least two parallel trajectories of
each patient's hemisphere. All patients provided written prior consent
for the procedure. The study was approved by the institutional review
board.
2.2. Operative procedure
For detailed description of our targeting approach and operative
procedure we refer to previously published reports of our group
(Nowacki et al., 2017). In summary, each patient underwent pre-
operative MRI to target the STN. Imaging was performed with a 12-
channel head coil for signal reception on a 3 T MRI system (MAGNE-
TOM TrioTim, Siemens, Germany). Imaging included T2-weighted
sequences with an echo time of 380 ms, a repetition time of 3000 ms
and a slice thickness of 1 mm. Targets and trajectories were identied
using Brainlab Elements software (Brainlab AG, Germany). On the day
of surgery a Leksell G frame (Elekta instruments, Sweden) was placed
and a high-resolution, stereotactic CT scan was performed and co-re-
gistered with the preoperative MRI (Brainlab AG, Germany).
Intraoperative microelectrode recording (MER) was performed as
Fig. 1. Determination of STN lengths in an example patient set. (A) Image-based measurements with Lead DBS and microelectrode tool that places microelectrodes in
specic trajectories and allows for measurement of STN length (blue points representing entry and exit, entry point for central trajectory covered by lead). The panel
shows the DISTAL atlas. GPe external globus pallidus, GPi internal globus pallidus, STN subthalamic nucleus, RN red nucleus. (B) Intraoperative assessment of
electrophysiological recording at our centre. Number of +signs indicate STN activity rated by one experienced rater live intraoperatively at dierent positions of
the microelectrode.
A. Nowacki et al. NeuroImage: Clinical 20 (2018) 868–874
869
described previously. Here, recording was performed on two to three
parallel trajectories. From 10 mm to 5 mm above the target, electro-
physiological activity was recorded in 1 mm-steps and from 5 mm
above the target in 0.5 mm-steps. Electrophysiological activity was
evaluated by a simple grading system, +for little activity to +++
for strong activity, to determine entry and exit of the STN (Fig. 1). This
grading was done live intraoperatively by an experienced rater and not
ad-hoc nor in a blinded fashion.
Postoperative high-resolution CT was performed for assessment of
correct electrode placement on the day after surgery.
2.3. Postoperative DBS lead location
2.3.1. Native space
Postoperative DBS lead localisation was performed based on
Brainlab Elements software in patient's native space. AC-PC coordinates
of the intended target, the DBS lead tip as well as the targeting error
were calculated as described previously (Nowacki et al., 2015). The
integrated workow was used to fuse preoperative MR images and post-
operative CT images. The STN was manually segmented by the junior
(AN) and senior neurosurgeon (CP) according to the hypointense signal
based on T2-weighted images. The DBS lead was reconstructed ac-
cording to the CT artefact in an automated way by the Brainlab soft-
ware and its position manually adjusted, if necessary.
2.3.2. Normalized Template space
Lead-DBS toolbox version 2.0.0.6 was used within Matlab 2016b
(The MathWorks, USA) for DBS lead visualisation (Horn and Kuhn,
2015). Preoperative MRI and postoperative CT were co-registered using
SPM12 and Advanced Normalization Tools (Avants et al., 2008). Nor-
malization to the MNI 152 2009b space (Montreal Neurological In-
stitute) was also performed by applying Advanced Normalization Tools.
The algorithm's accuracy and eectiveness were evaluated elsewhere
(Klein et al., 2009). The lead was then automatically pre-localised on
CT images with PaCER and manually adjusted, if the automatic loca-
lisation did not match the CT artefact (typical adjustments of less than
half of the lead's diameter or 0.65 mm) (Husch et al., 2018). Lead-DBS
provided a three-dimensional display of nuclei according to a selected
atlas. Four atlases were evaluated: the DISTAL atlas combining MRI
from the 2009b MNI152 template (normative young adult population),
histology and structural connectivity data (Ewert et al., 2017); the
CIT168 atlas derived from data of the Human Connectome Project
based on 168 young adult subjects between the age of 22 and 35 (Pauli
et al., 2017); the Wang atlas based on high-resolution 7 T MR scans of
twelve healthy subjects; and the PD25 atlas derived from 3 T MRI scans
of 25 Parkinson's disease patients (Xiao et al., 2015). Our rationale
behind the selection of these four atlases was to cover a broad spectrum
of modalities, e.g., dierent MRI data, histology and subject groups.
2.4. Anatomical comparison of STN in native and template space
To assess the accuracy of the four dierent brain atlases, we com-
pared each patient's individual STN anatomy based on the hypointense
T2-artefact on their MRI and the postoperative DBS lead location (na-
tive space) with the reconstructed DBS lead and STN according to the
four brain atlases after normalization to the MNI space (template
space).
Furthermore, we compared the volumes of each patient's in-
dividually segmented STN (left and right) according to the MRI-based
T2 artefact and the atlas-based STN volume (supplementary g. 1). In
Lead-DBS, the segmented STNs according to the four atlases were ex-
ported and then analysed in Matlab 2016b to compute the volume.
The Dice coecient was applied as a measure of the degree of si-
milarity between the native space STN and the atlas-based STN. It was
calculated according to Eq. (1), where A and B are the STN volumes in
native and template space, respectively.
=∣∩∣
+∣ ∣
Dice AB
AB
2
|| (1)
2.5. STN MER-MRI trajectory length comparison
One further pertinent way to assess the accuracy of the re-
constructed DBS lead relative to STN anatomy was to compare the
position of the DBS lead with intraoperative MER results. Therefore, we
determined the length of electrophysiological STN activity along the
two to three intraoperatively used parallel trajectories and compared it
to image-based reconstructed trajectories across the STN according to
the four atlases (template space) as well as to the T2-hypointense signal
of each patient's STN visualised on MRI (native space).
Native-space based analysis was performed with Brainlab software.
Reconstructed trajectories of MER adjacent to the DBS lead were
manually placed in inline view with a xed 2 mm distance to the lead
according to known features of the ve-channel Ben-Gun. The entry and
exit points of the STN were then measured manually in the software to
calculate the resulting STN length. Lead-DBS provided a microelectrode
tool to determine STN lengths along the dierent microelectrode tra-
jectories such as central, lateral or medial (Fig. 1).
We then calculated the dierence of the image- or atlas-based STN
trajectory length (S
MRI/atlas
) and the MER-based trajectory length (S
MER
)
in each patient (ΔS=S
MRI/atlas
-S
MER
). If no electrophysiological ac-
tivity was recorded for a given trajectory, S
MER
was set to zero.
Similarly, if the image- or atlas-based trajectory was projected outside
the STN, S
MRI/atlas
was set to zero for that trajectory.
2.6. Clinical outcome assessment
Preoperative motor subscores of the Unied Parkinson's Disease
Rating Scale (UPDRS part III) in medication-ostate and post-
operative UPDRS III scores in the stimulation-on, medication-ostate
(overnight withdrawal of PD medication) were systematically de-
termined 12 months after DBS implantation. DBS outcome was assessed
by analysing the dierences of respective preoperative and post-
operative UPDRS III scores.
2.7. Statistical analysis
For both anatomical and trajectory length comparisons, measure-
ments were done separately for the left and right hemispheres to ac-
count for asymmetries. Finally, data were then analysed together with
descriptive/nonparametric statistics using Prism software (GraphPad
Prism 6, USA). One-way ANOVA (Friedman test) was applied to test for
statistical dierences between the image-based and electro-
physiological-based STN lengths. Post-hoc analysis was performed by
Dunn's multiple comparisons test. A P-value < .05 was considered
statistically signicant.
3. Results
According to our inclusion criteria we included a total number of
ten patients (19 hemispheres) in the nal analysis. Altogether, a total
number of 47 MER trajectories through the STN were analysed (cf.
Table I, 10 along a medial trajectory, 19 along a central trajectory, 18
along a lateral trajectory). The mean targeting error was
1.01 ± 0.57 mm (range 0.401.95 mm) indicating precise targeting
and no cases of clinically relevant lead displacement.
3.1. Clinical outcome data
Preoperative mean UPDRS III score was 36.9 ± 9.75. Postoperative
mean UPDRS III score was signicantly reduced to 10.55 ± 6.40
(p< .0001). The mean percentage improvement was
A. Nowacki et al. NeuroImage: Clinical 20 (2018) 868–874
870
71.72 ± 14.96%. One patient was lost to clinical follow-up as he died
(unrelated to DBS surgery).
3.2. DBS lead relative to STN anatomy in native and template space
The mean AC-PC coordinates of the right lead tip were LAT
11.40 ± 1.10 mm, ANT -3.31 ± 0.39 mm and VERT
-4.43 ± 0.55 mm. After transformation of the DBS leads into the MNI
space the mean MNI coordinates were LAT 9.38 ± 1.33 mm, AP
-4.44 ± 0.83 mm and VERT -9.35 ± 1.2 mm. Table 1 summarizes
patients´ individual DBS lead tip position relative to the MCP in native
space and after normalization in the MNI space.
Electrode reconstruction after normalization and application of the
four dierent atlases led to dierent results with respect to the lead
position within the STN and compared to the native space. Fig. 2 vi-
sualises the DBS lead and STN delineation of the manually segmented
STN compared to the dierent atlases in one representative patient of
the study cohort. In general, we could observe that the DBS lead was
projected more laterally within the STN after applying the DISTAL and
CIT atlas in comparison to the native space. No such shift could be
observed in case of the PD25 and Wang atlases, in which cases, how-
ever, the DBS lead was frequently located outside the atlas-based STN
borders.
The mean STN volume of the study cohort was 153.3 ± 40.3 mm
3
(n= 19). This is similar to the STN volume of the DISTAL atlas
(166 mm
3
;p= .22), but signicantly larger compared to the CIT atlas
(136 mm
3
;p= .04), the atlas of Wang et al. (85 mm
3
;p< .001) and
the PD25 atlas (102 mm
3
; p < .001).
The Dice coecient of the native space STN volume compared to
each of the dierent atlas-based STN was determined in each patient's
hemisphere. The overlap was highest for the DISTAL atlas
(0.56 ± 0.18). Less similar results were found for the CIT atlas
(0.50 ± 0.15). Furthermore, we found signicant dierences for the
atlas of Wang et al. (0.42 ± 0.12) as well as the PD25 (0.34 ± 0.17).
3.3. STN MER-MRI trajectory comparison in native space and template
space
Based on all three trajectories, there was no signicant dierence
between ΔS in native space, however there was a statistically signicant
dierence of ΔS for all four atlases (p< .0001).
Fig. 3 shows the dierences of the electrophysiological and image-
based STN trajectory lengths separately for the central, medial and
lateral trajectories.
The mean dierences (ΔS) between the reconstructed MRI-based
trajectory length in native space and intraoperatively measured MER-
based STN length was 0.49 ± 0.91 mm for the central trajectory,
0.45 ± 2.10 mm for the medial and 0.97 ± 2.99 mm for the lateral
trajectory.
The mean dierences (ΔS) between the reconstructed DISTAL atlas-
based STN length and MER-based STN length was 0.17 ± 1.68 mm
for the central trajectory, 0.33 ± 2.75 mm for the medial
Table 1
Coordinates of right lead tip in native and MNI space with reference to the mid-commissural point (MCP).
ID MCP Coordinates in native space MCP coordinates in MNI space
LAT AP VERT LAT AP VERT
1 9.49 3.73 5.73 7.49 3.68 10.57
2 11.47 3.63 4.59 8.44 5.82 10.73
3 12.93 3.65 3.98 11.17 5.2 7.33
4 11.74 3.51 4 9.95 4.4 8.44
5 12.02 3.03 4.12 10.72 4.28 8.14
6 12.38 3.41 4.4 10.02 5.05 10.72
7 10.43 2.6 4.78 9.67 3.35 9.83
8 12.04 3.6 3.92 9.18 4.26 9.17
9 10.03 2.91 4.67 7.11 3.36 10.02
10 11.51 3.01 4.15 10.09 5.06 8.55
mean 11.40 3.31 4.43 9.38 4.44 9.35
SD 1.10 0.39 0.55 1.33 0.83 1.20
Fig. 2. Lead position in relation to STN. (A) Postoperative CT lead artefact (red) superimposed on preoperative MR T2 images. (B)-(E) Right lead marked as asterisk
in atlas segmented STN for DISTAL (B), CIT (C), Wang (D) and PD25 (E), respectively.
A. Nowacki et al. NeuroImage: Clinical 20 (2018) 868–874
871
and 2.58 ± 2.40 mm for the lateral trajectory. Frequently, the lat-
eral microelectrode trajectory was displayed outside the DISTAL atlas-
based STN, while there was marked electrophysiological activity re-
corded intraoperatively. This led to negative ΔS-values on the lateral
trajectory in those cases. On the other hand medial trajectories were
displayed within the DISTAL atlas-based STN, even though no electro-
physiological activity was measured (Fig. 4).
The mean dierences (ΔS) between the reconstructed CIT atlas-
based STN length and MER-based STN length was 0.93 ± 1.50 mm
for the central trajectory, 1.13 ± 1.99 mm for the medial
and 2.67 ± 2.21 mm for the lateral trajectory.
The mean dierences (ΔS) between the reconstructed Wang atlas-
based STN length and MER-based STN length was 2.42 ± 1.98 mm
for the central trajectory, 0.7 ± 2.56 mm for the medial
and 2.71 ± 2.28 mm for the lateral trajectory.
The mean dierences (ΔS) between the reconstructed PD25 atlas-
based STN length and MER-based STN length was 1.62 ± 1.90 mm
for the central trajectory, 1.45 ± 2.38 mm for the medial
and 2.94 ± 2.05 mm for the lateral trajectory.
These dierence calculations were repeated in native space by
transforming the four atlases from MNI space to native patient space.
There we also observed the smallest dierences for the DISTAL atlas
(supplementary g. 2).
Fig. 3. Dierences in measured STN lengths between manually segmented, atlas based STN and intraoperative electrophysiology data. Thick red lines indicate mean;
red boxes represent the 95% condence interval of the mean; blue box represents the standard deviation; gray dots represent individual data.
Fig. 4. Overlay of manually segmented STN (orange) and atlas based STN according to Distal (green), CIT (yellow), Wang (rose) and PD25 (gray) of an example
patient. The lead tip was used as anchor to align manually segmented and atlas-based STNs.
A. Nowacki et al. NeuroImage: Clinical 20 (2018) 868–874
872
4. Discussion
Our study results show that lead reconstruction after normalization
into the MNI template space and application of four dierent atlases led
to dierent results in terms of the DBS lead position relative to the STN.
The STN volume and delineation of the four dierent atlases as well as
the relation to the reconstructed DBS lead were compared to the T2-
hypointense MRI signal in each individual patient and to the STN
electrophysiological prole recorded during intraoperative MER. We
demonstrated that the process of normalization into the MNI template
space and application of any of the four tested atlases resulted in sig-
nicant dierences between the electrophysiological prole and the
atlas-based prole of the STN. Furthermore, signicant dierences were
found between atlases regarding volume and similarity index compared
to native space. Based on electrophysiological and imaging data, the
DISTAL atlas led to the most accurate display of the reconstructed DBS
lead relative to the DISTAL-based STN. Notably, after applying the
DISTAL atlas the reconstructed DBS leads were displayed too lateral
within the STN compared to native space and to the electro-
physiological prole.
Accurate display of the DBS lead in relation to the surrounding
anatomical structures is of paramount importance in clinical routine as
well as for scientic purposes such as cohort comparisons.
Understanding the relation of the lead to the surrounding anatomical
structures may help clinicians to adapt stimulation parameters to avoid
side eects and optimize stimulation. Importantly, given the fact that
the target structures for DBS are usually small in size with only a few
millimetres of diameter, accurate DBS lead reconstruction is crucial
when it comes to comparison of DBS lead placements across subjects to
infer optimal stimulation sites (Horn et al., 2017). Three-dimensional
subcortical human brain atlases (Cheung et al., 2014;Welter et al.,
2014) in conjunction with computational models of volume of tissue
activated or bre tractography are helpful tools to address these
questions (Gunalan et al., 2018). Dierent subcortical atlases have been
developed to enable microstructural analysis of brain anatomy. Some
atlases are based on histologic data of single or multiple brains, while
others are dened solely through MRI data. Some more recent atlases,
such as the DISTAL atlas studied herein, combine several modalities to
improve accuracy (Ewert et al., 2018).
From each patient's data set, we concluded that the DBS lead arte-
fact on postoperative CT reected the actual lead position within the
STN reliably as we found no dierence between the reconstructed
image-based trajectories through the STN and their corresponding
electrophysiological activity. This is in line with previous work by our
and other groups (Hamani et al., 2005;Kocabicak et al., 2013;Nowacki
et al., 2017;Schlaier et al., 2011).
On the other hand, normalization of individual MRI patient data
into the MNI space and application of the four atlases tested inevitably
led to certain shifts of the reconstructed DBS lead in relation to the STN
anatomy. One factor is the normalization to the MNI space. A com-
parison of patients lead tip position with reference to MCP in native
space and after normalization into the MNI template space shows a
dierence in all coordinates especially in the vertical direction. A
thorough discussion and probabilistic conversion of AC/PC coordinates
to MNI coordinates is provided by Horn et al. (Horn et al., 2017). In
addition, MNI space denes the origin close to the AC point, not quite
on the AC-PC line, but a few millimetres more inferior.
Comparison between atlas-based reconstructed MER trajectories
through the STN showed signicant dierences compared to the elec-
trophysiological activity. However, the degree of discrepancy between
imaging and electrophysiological data was considerably dierent be-
tween the four atlases tested. Whereas there was a good correspondence
of electrophysiological activity and the reconstructed MER trajectory
through the STN of the DISTAL and CIT atlases along the central tra-
jectory, there was a low correspondence in case of the Wang and PD25
atlas along each of the tested trajectories. It is important to note that,
frequently, lateral trajectories were displayed outside the DISTAL and
CIT atlas STN, while there was marked electrophysiological activity
leading to negative ΔS-values. On the other hand, medial trajectories
were displayed within the STN even though no electrophysiological
activity was measured. It highlights the tendency of a lateral shift of the
reconstructed DBS lead relative to all four atlas-based STN. This shift
might be due to the normalization algorithm, specically as the algo-
rithm transforms brain images of Parkinson's disease patients which
typically show ventricular enlargement but smaller basal ganglia
structures than normal controls to the MNI template brain (Geng
et al., 2006). This warrants further investigation in order to adapt the
normalization algorithm for DBS research.
Our data further showed that the DISTAL atlas-based STN volume
corresponded well to the mean STN volume in our cohort of patients.
Application of the other atlases led to an underestimation of the STN
volume, despite the fact that the volume of the MNI template brain
tends to be larger than the average human brain (Horn et al., 2017).
Similarity analysis revealed the best correspondence between the STN
shape of the DISTAL atlas compared to each patient's individual STN.
Poor anatomical overlap was found between the Wang and PD25 atlas
STN and our cohort of patients.
Our study has some limitations. First, our relatively small sample
size of ten patients that we have included in the nal analysis prevents
generalizing our ndings to a broader population. Second, we only
analysed four dierent atlases. Literature now abounds with dierent
subcortical brain atlases. Analysing all of them was beyond the scope of
this study. We focused on the four presented atlases here to compare
and contrast the multimodality of the DISTAL atlas, the high resolution
and high subject count (168) of the CIT atlas, the 7 T based Wang atlas
and the disease-specic PD25 atlas based on Parkinson's disease pa-
tients. While the DISTAL atlas was specically dened in the MNI space
used herein, the other three atlases were each dened on custom
templates. The CIT atlas was based on data from the Human
Connectome Project, the 7 T based atlas from a groupwise template
built from the data of the twelve healthy subjects scanned and the PD25
atlas was dened as an averaged atlas across the 25 Parkinson's disease
patients. These three atlases were transformed to the MNI template
space and the transformed registrations were provided within LEAD-
DBS. This additional transformation may have aected the comparison.
However, for better comparison we specically intended to analyse all
four atlases in the same template space. Third, analysis of the STN size
and shape in native space was based on manual segmentation of the T2-
hypointense STN signal based on MRI ignoring the fact that dierent
MRI sequences have been proposed to display the STN, such as sus-
ceptibility weighted sequences or uid attenuated inversion recovery
sequences (Chandran et al., 2016;Heo et al., 2015;Senova et al., 2016).
Furthermore, delineation of the STN borders by imaging is not always
clear, which has been described previously by our and other groups
(Nowacki et al., 2017;Starr et al., 2002). Especially the anterior-pos-
terior axis of the STN was found to be prone to MRI distortions
(Sumanaweera et al., 1994). Similarly, delineation of the STN with
electrophysiological data was rated live intraoperatively by a single
experienced rater and not in a blinded retrospective fashion with
multiple raters or with an automised algorithm (Zaidel et al., 2010).
Fourth, our analysis focused on STN length and did not study entry and
exit points in detail that could be prone to error by dierent lead lo-
calization algorithms and manual corrections. Finally, we relied on
previous evaluation of registration and normalization algorithms
15
, but
we think that further improvements in this domain may be warranted.
Furthermore, we performed an early post-operative CT scan as part of
the clinical routine to rule out complications. Due to resorption of
possible postoperative pneumencephalus, the lead position might
slightly change over the postoperative period and introduce a potential
bias into DBS lead position analysis. Furthermore, we did not measure
the electrophysiological activity along an anterior or posterior trajec-
tory, which would have added great value to evaluate the anatomical
A. Nowacki et al. NeuroImage: Clinical 20 (2018) 868–874
873
accuracy of the atlases with respect to the border between the anterior
STN and internal capsule.
5. Conclusions
Normalization of individual MRI data into the MNI template space
and application of dierent three-dimensional subcortical atlases in-
evitably led to distortions of the reconstructed lead in relation to the
STN. Of the four dierent atlases tested in this work, the DISTAL atlas
leads to the most accurate results. However, results must be cautiously
interpreted as there seems to be a lateral shift of the reconstructed DBS
lead relative to the STN.
Funding
This research did not receive any specic grant from funding
agencies in the public, commercial, or not-for-prot sectors.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.nicl.2018.09.030.
References
Accolla, E.A., Herrojo Ruiz, M., Horn, A., Schneider, G.H., Schmitz-Hubsch, T., Draganski,
B., Kuhn, A.A., 2016. Brain networks modulated by subthalamic nucleus deep brain
stimulation. Brain 139, 25032515.
Akram, H., Sotiropoulos, S.N., Jbabdi, S., Georgiev, D., Mahlknecht, P., Hyam, J.,
Foltynie, T., Limousin, P., De Vita, E., Jahanshahi, M., Hariz, M., Ashburner, J.,
Behrens, T., Zrinzo, L., 2017. Subthalamic deep brain stimulation sweet spots and
hyperdirect cortical connectivity in Parkinson's disease. Neuroimage 158, 332345.
Avants, B.B., Epstein, C.L., Grossman, M., Gee, J.C., 2008. Symmetric dieomorphic
image registration with cross-correlation: evaluating automated labeling of elderly
and neurodegenerative brain. Med Image Anal 12, 2641.
Brett, M., Johnsrude, I.S., Owen, A.M., 2002. The problem of functional localization in the
human brain. Nat Rev. Neurosci 3, 243249.
Chakravarty, M.M., Bertrand, G., Hodge, C.P., Sadikot, A.F., Collins, D.L., 2006. The
creation of a brain atlas for image guided neurosurgery using serial histological data.
Neuroimage 30, 359376.
Chandran, A.S., Bynevelt, M., Lind, C.R., 2016. Magnetic resonance imaging of the sub-
thalamic nucleus for deep brain stimulation. J Neurosurg 124, 96105.
Cheung, T., Noecker, A.M., Alterman, R.L., McIntyre, C.C., Tagliati, M., 2014. Dening a
therapeutic target for pallidal deep brain stimulation for dystonia. Ann Neurol 76,
2230.
Dalaker, T.O., Zivadinov, R., Ramasamy, D.P., Beyer, M.K., Alves, G., Bronnick, K.S.,
Tysnes, O.B., Aarsland, D., Larsen, J.P., 2011. Ventricular enlargement and mild
cognitive impairment in early Parkinson's disease. Mov Disord 26, 297301.
Deuschl, G., Schade-Brittinger, C., Krack, P., Volkmann, J., Schafer, H., Botzel, K.,
Daniels, C., Deutschlander, A., Dillmann, U., Eisner, W., Gruber, D., Hamel, W.,
Herzog, J., Hilker, R., Klebe, S., Kloss, M., Koy, J., Krause, M., Kupsch, A., Lorenz, D.,
Lorenzl, S., Mehdorn, H.M., Moringlane, J.R., Oertel, W., Pinsker, M.O., Reichmann,
H., Reuss, A., Schneider, G.H., Schnitzler, A., Steude, U., Sturm, V., Timmermann, L.,
Tronnier, V., Trottenberg, T., Wojtecki, L., Wolf, E., Poewe, W., Voges, J., German
Parkinson Study Group, N.S, 2006. A randomized trial of deep-brain stimulation for
Parkinson's disease. N Engl J Med 355, 896908.
Ewert, S., Plettig, P., Li, N., Chakravarty, M.M., Collins, D.L., Herrington, T.M., Kuhn,
A.A., Horn, A., 2017. Toward dening deep brain stimulation targets in MNI space: A
subcortical atlas based on multimodal MRI, histology and structural connectivity.
Neuroimage. 170, 271282.
Ewert, S., Plettig, P., Li, N., Chakravarty, M.M., Collins, D.L., Herrington, T.M., Kuhn,
A.A., Horn, A., 2018. Toward dening deep brain stimulation targets in MNI space: A
subcortical atlas based on multimodal MRI, histology and structural connectivity.
Neuroimage 170, 271282.
Geng, D.Y., Li, Y.X., Zee, C.S., 2006. Magnetic resonance imaging-based volumetric
analysis of basal ganglia nuclei and substantia nigra in patients with Parkinson's
disease. Neurosurgery 58, 256262.
Gunalan, K., Howell, B., McIntyre, C.C., 2018. Quantifying axonal responses in patient-
specic models of subthalamic deep brain stimulation. Neuroimage 172, 263277.
Hamani, C., Richter, E.O., Andrade-Souza, Y., Hutchison, W., Saint-Cyr, J.A., Lozano,
A.M., 2005. Correspondence of microelectrode mapping with magnetic resonance
imaging for subthalamic nucleus procedures. Surg Neurol 63, 249253 discussion
253.
Heo, Y.J., Kim, S.J., Kim, H.S., Choi, C.G., Jung, S.C., Lee, J.K., Lee, C.S., Chung, S.J., Cho,
S.H., Lee, G.R., 2015. Three-dimensional uid-attenuated inversion recovery se-
quence for visualisation of subthalamic nucleus for deep brain stimulation in
Parkinson's disease. Neuroradiology 57, 929935.
Horn, A., Kuhn, A.A., 2015. Lead-DBS: a toolbox for deep brain stimulation electrode
localizations and visualizations. Neuroimage 107, 127135.
Horn, A., Kuhn, A.A., Merkl, A., Shih, L., Alterman, R., Fox, M., 2017. Probabilistic
conversion of neurosurgical DBS electrode coordinates into MNI space. Neuroimage
150, 395404.
Husch, A., M, V.P., Gemmar, P., Goncalves, J., Hertel, F., 2018. PaCER - A fully automated
method for electrode trajectory and contact reconstruction in deep brain stimulation.
Neuroimage Clin 17, 8089.
Klein, A., Andersson, J., Ardekani, B.A., Ashburner, J., Avants, B., Chiang, M.C.,
Christensen, G.E., Collins, D.L., Gee, J., Hellier, P., Song, J.H., Jenkinson, M., Lepage,
C., Rueckert, D., Thompson, P., Vercauteren, T., Woods, R.P., Mann, J.J., Parsey,
R.V., 2009. Evaluation of 14 nonlinear deformation algorithms applied to human
brain MRI registration. Neuroimage 46, 786802.
Kocabicak, E., Aygun, D., Ozaydin, I., Jahanshahi, A., Tan, S., Onar, M., Boke, O., Kurt,
M., Guz, H., Terzi, M., Alptekin, O., Temel, Y., 2013. Does probe's eye subthalamic
nucleus length on T2W MRI correspond with microelectrode recording in patients
with deep brain stimulation for advanced Parkinson's disease? Turk Neurosurg 23,
658665.
Nowacki, A., Fiechter, M., Fichtner, J., Debove, I., Lachenmayer, L., Schupbach, M.,
Oertel, M.F., Wiest, R., Pollo, C., 2015. Using MDEFT MRI Sequences to Target the
GPi in DBS Surgery. PLoS One 10, e0137868.
Nowacki, A., Debove, I., Fiechter, M., Rossi, F., Oertel, M.F., Wiest, R., Schupbach, M.,
Pollo, C., 2017. Targeting Accuracy of the Subthalamic Nucleus in Deep Brain
Stimulation Surgery: Comparison Between 3 T T2-Weighted Magnetic Resonance
Imaging and Microelectrode Recording Results. Oper Neurosurg (Hagerstown) 15 (1),
6671.
Pauli, W.M., Nili, A.N., Tyszka, J.M., 2017. A High-Resolution Probabilistic In Vivo Atlas
of Human Subcortical Brain Nuclei. bioRxiv 211,201.
Pollo, C., Kaelin-Lang, A., Oertel, M.F., Stieglitz, L., Taub, E., Fuhr, P., Lozano, A.M.,
Raabe, A., Schupbach, M., 2014. Directional deep brain stimulation: an in-
traoperative double-blind pilot study. Brain 137, 20152026.
Schlaier, J.R., Habermeyer, C., Warnat, J., Lange, M., Janzen, A., Hochreiter, A.,
Proescholdt, M., Brawanski, A., Fellner, C., 2011. Discrepancies between the MRI-
and the electrophysiologically dened subthalamic nucleus. Acta Neurochir (Wien)
153, 23072318.
Schuepbach, W.M., Rau, J., Knudsen, K., Volkmann, J., Krack, P., Timmermann, L.,
Halbig, T.D., Hesekamp, H., Navarro, S.M., Meier, N., Falk, D., Mehdorn, M.,
Paschen, S., Maarouf, M., Barbe, M.T., Fink, G.R., Kupsch, A., Gruber, D., Schneider,
G.H., Seigneuret, E., Kistner, A., Chaynes, P., Ory-Magne, F., Brefel Courbon, C.,
Vesper, J., Schnitzler, A., Wojtecki, L., Houeto, J.L., Bataille, B., Maltete, D., Damier,
P., Raoul, S., Sixel-Doering, F., Hellwig, D., Gharabaghi, A., Kruger, R., Pinsker, M.O.,
Amtage, F., Regis, J.M., Witjas, T., Thobois, S., Mertens, P., Kloss, M., Hartmann, A.,
Oertel, W.H., Post, B., Speelman, H., Agid, Y., Schade-Brittinger, C., Deuschl, G.,
Group, E.S, 2013. Neurostimulation for Parkinson's disease with early motor com-
plications. N Engl J Med 368, 610622.
Schupbach, W.M.M., Chabardes, S., Matthies, C., Pollo, C., Steigerwald, F., Timmermann,
L., Visser Vandewalle, V., Volkmann, J., Schuurman, P.R., 2017. Directional leads for
deep brain stimulation: Opportunities and challenges. Mov Disord 32, 13711375.
Senova, S., Hosomi, K., Gurruchaga, J.M., Gouello, G., Ouerchefani, N., Beaugendre, Y.,
Lepetit, H., Lefaucheur, J.P., Badin, R.A., Dauguet, J., Jan, C., Hantraye, P.,
Brugieres, P., Pal, S., 2016. Three-dimensional SPACE uid-attenuated inversion
recovery at 3 T to improve subthalamic nucleus lead placement for deep brain sti-
mulation in Parkinson's disease: from preclinical to clinical studies. J Neurosurg 125,
472480.
Starr, P.A., Christine, C.W., Theodosopoulos, P.V., Lindsey, N., Byrd, D., Mosley, A.,
Marks Jr., W.J., 2002. Implantation of deep brain stimulators into the subthalamic
nucleus: technical approach and magnetic resonance imaging-veried lead locations.
J Neurosurg 97, 370387.
Sumanaweera, T.S., Adler, J.R., Jr., Napel, S., Glover, G.H., 1994. Characterization of
spatial distortion in magnetic resonance imaging and its implications for stereotactic
surgery. Neurosurgery 35, 696703; discussion 703694.
Tinkhauser, G., Pogosyan, A., Debove, I., Nowacki, A., Shah, S.A., Seidel, K., Tan, H.,
Brittain, J.S., Petermann, K., di Biase, L., Oertel, M., Pollo, C., Brown, P., Schuepbach,
M., 2018. Directional local eld potentials: A tool to optimize deep brain stimulation.
Mov Disord 33, 159164.
Wang, B.T., Poirier, S., Guo, T., Parrent, A.G., Peters, T.M., Khan, A.R., 2016. Generation
and evaluation of an ultra-high-eld atlas with applications in DBS planning. SPIE
Medical Imaging. SPIE, pp. 10.
Welter, M.L., Schupbach, M., Czernecki, V., Karachi, C., Fernandez-Vidal, S., Golmard,
J.L., Serra, G., Navarro, S., Welaratne, A., Hartmann, A., Mesnage, V., Pineau, F.,
Cornu, P., Pidoux, B., Worbe, Y., Zikos, P., Grabli, D., Galanaud, D., Bonnet, A.M.,
Belaid, H., Dormont, D., Vidailhet, M., Mallet, L., Houeto, J.L., Bardinet, E., Yelnik, J.,
Agid, Y., 2014. Optimal target localization for subthalamic stimulation in patients
with Parkinson disease. Neurology 82, 13521361.
Wodarg, F., Herzog, J., Reese, R., Falk, D., Pinsker, M.O., Steigerwald, F., Jansen, O.,
Deuschl, G., Mehdorn, H.M., Volkmann, J., 2012. Stimulation site within the MRI-
dened STN predicts postoperative motor outcome. Mov Disord 27, 874879.
Xiao, Y., Fonov, V., Beriault, S., Al Subaie, F., Chakravarty, M.M., Sadikot, A.F., Pike,
G.B., Collins, D.L., 2015. Multi-contrast unbiased MRI atlas of a Parkinson's disease
population. Int J Comput Assist Radiol Surg 10, 329341.
Yelnik, J., Bardinet, E., Dormont, D., Malandain, G., Ourselin, S., Tande, D., Karachi, C.,
Ayache, N., Cornu, P., Agid, Y., 2007. A three-dimensional, histological and de-
formable atlas of the human basal ganglia. I. Atlas construction based on im-
munohistochemical and MRI data. Neuroimage 34, 618638.
Zaidel, A., Spivak, A., Grieb, B., Bergman, H., Israel, Z., 2010. Subthalamic span of beta
oscillations predicts deep brain stimulation ecacy for patients with Parkinson's
disease. Brain 133, 20072021.
A. Nowacki et al. NeuroImage: Clinical 20 (2018) 868–874
874

Supplementary resource (1)

... For this seed region re-analysis, we used the Harvard Ascending Arousal Network atlas [36,37] in the Lead DBS software package [38][39][40] to define regions of interest (ROI) for midbrain nuclei. The mean BOLD signal for each ROI was extracted from each subject's contrast map, re-centered to a population grand mean of 0, and the normalized data were analyzed in the MarsBaR 0.44 toolbox [41,42]. ...
... The findings were inferred based on the seed regions extracted from the ascending arousal network. Coordinates of the seed regions may improve as newer standards are created [40]. The actual metric being compared is the 2-back>0-back differential activity during the difficult cognitive working memory task. ...
Article
Full-text available
Background: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS), Gulf War Illness (GWI) and control subjects underwent fMRI during difficult cognitive tests performed before and after submaximal exercise provocation (Washington 2020). Exercise caused increased activation in ME/CFS but decreased activation for GWI in the dorsal midbrain, left Rolandic operculum and right middle insula. Midbrain and isthmus nuclei participate in threat assessment, attention, cognition, mood, pain, sleep, and autonomic dysfunction. Methods: Activated midbrain nuclei were inferred by a re-analysis of data from 31 control, 36 ME/CFS and 78 GWI subjects using a seed region approach and the Harvard Ascending Arousal Network. Results: Before exercise, control and GWI subjects showed greater activation during cognition than ME/CFS in the left pedunculotegmental nucleus. Post exercise, ME/CFS subjects showed greater activation than GWI ones for midline periaqueductal gray, dorsal and median raphe, and right midbrain reticular formation, parabrachial complex and locus coeruleus. The change between days (delta) was positive for ME/CFS but negative for GWI, indicating reciprocal patterns of activation. The controls had no changes. Conclusions: Exercise caused the opposite effects with increased activation in ME/CFS but decreased activation in GWI, indicating different pathophysiological responses to exertion and mechanisms of disease. Midbrain and isthmus nuclei contribute to postexertional malaise in ME/CFS and GWI.
... Thus, general human applications have been first thoroughly studied using animal models, even in those cases of testing trials using volunteer human subjects. By way of example, the cases in [87] and [88] can be cited as cases of human experimental interventions. Secondly, when considering research applications using animal models, where some cognitive hypotheses, e.g. ...
Article
Full-text available
Modifying neural activity is a substantial goal in neuroscience that facilitates the understanding of brain functions and the development of medical therapies. Neurobiological models play an essential role, contributing to the understanding of the underlying brain dynamics. In this context, control systems represent a fundamental tool to provide a correct articulation between model stimulus (system inputs) and outcomes (system outputs). However, throughout the literature there is a lack of discussions on neurobiological models, from the formal control perspective. In general, existing control proposals applied to this family of systems, are developed empirically, without theoretical and rigorous framework. Thus, the existing control solutions, present clear and significant limitations. The focus of this work is to survey dynamical neurobiological models that could serve for closed-loop control schemes or for simulation analysis. Consequently, this paper provides a comprehensive guide to discuss and analyze control-oriented neurobiological models. It also provides a potential framework to adequately tackle control problems that could modify the behavior of single neurons or networks. Thus, this study constitutes a key element in the upcoming discussions and studies regarding control methodologies applied to neurobiological systems, to extend the present research and understanding horizon for this field.
... However, it is unclear how well they reflect individual anatomy, particularly in the context of a progressive neurodegenerative disease [32]. Other limitations of image processing pipelines include those related to registration (which may be ameliorated to a degree by post-operative MRI [33], and the utilisation of non-automated processes in electrode reconstruction (which is currently easier with post-operative CT and may also potentially allow directionality determination [20,34]. ...
Preprint
Full-text available
Introduction Accurate placement of deep brain stimulation electrodes within the intended target is believed to be a key variable related to outcomes. However, methods to verify electrode location are not universally established. Research Question The aim of this study was to determine the feasibility of post-op lead localisation in clinical practice and its utility to audit our own DBS accuracy. Material and Methods A retrospective cohort study was performed of a consecutive series of patients with Parkinson’s disease who underwent deep brain stimulation of either the globus pallidus internus (GPi) or subthalamic nucleus (STN) between 2016 and 2019. Image processing was performed using the Lead-DBS toolbox. Institutional ethical approval was granted as a review of service. Results In total 38 participants met the inclusion criteria. Electrode localisation was completed in 79%. Clinical outcomes included improvement in UPDRS III of 46% and PDQ39 of 32%. Overall electrode accuracy was 0.22 +/- 0.4mm for all electrodes to the main nucleus with 9 (12%) outliers but only 3 (4%) electrodes out with 2mm from the intended target. Accuracy was worse for the second electrode implanted and in the GPi but was not affected by pneumocephalus or brain shift. Neither clinical outcomes nor the volume of activated tissue was affected by electrode accuracy. Discussion and Conclusions A neuroimaging approach to electrode localisation allows qualitative appraisal of targeting accuracy and is feasible with routine clinical data. Such methods are complimentary to traditional co-ordinate approaches and lend themselves to developing large, collaborative, and quantitative projects. HIGHLIGHTS Overall electrode accuracy was 0.22 +/- 0.4mm with only 3 (4%) electrodes out with 2mm from the intended target Accuracy was significantly worse in the GPi versus the STN and on the second side implanted Inaccuracy occurred in the X (lateral) plane but was not related to pneumocephalus or brain shift
... Recently, Nowacki et al. (Nowacki et al., 2018) analyzed the accuracy of DBS electrode locations using atlases provided within the Lead-DBS package: in all regards, the DISTAL atlas (Ewert et al., 2018), which was used here, was the most accurate atlas. However, this group showed that, although very close, the STN length was significantly different when comparing the measurement performed using the DISTAL atlas and the one obtained from MERs. ...
Article
Full-text available
Microelectrode recordings (MERs) are often used during deep brain stimulation (DBS) surgeries to confirm the position of electrodes in patients with advanced Parkinson’s disease. The present study focused on 32 patients who had undergone DBS surgery for advanced Parkinson’s disease. The first objective was to confront the anatomical locations of intraoperative individual MERs as determined electrophysiologically with those determined postoperatively by image reconstructions. The second aim was to search for differences in cell characteristics among the three subthalamic nucleus (STN) subdivisions and between the STN and other identified subcortical structures. Using the DISTAL atlas implemented in the Lead-DBS image reconstruction toolbox, each MER location was determined postoperatively and attributed to specific anatomical structures (sensorimotor, associative or limbic STN; substantia nigra [SN], thalamus, nucleus reticularis polaris, zona incerta [ZI]). The STN dorsal borders determined intraoperatively from electrophysiology were then compared with the STN dorsal borders determined by the reconstructed images. Parameters of spike clusters (firing rates, amplitudes – with minimum amplitude of 60 μV -, spike durations, amplitude spectral density of β-oscillations) were compared between structures (ANOVAs on ranks). Two hundred and thirty one MERs were analyzed (144 in 34 STNs, 7 in 4 thalami, 5 in 4 ZIs, 34 in 10 SNs, 41 others). The average difference in depth of the electrophysiological dorsal STN entry in comparison with the STN entry obtained with Lead-DBS was found to be of 0.1 mm (standard deviation: 0.8 mm). All 12 analyzed MERs recorded above the electrophysiologically-determined STN entry were confirmed to be in the thalamus or zona incerta. All MERs electrophysiologically attributed to the SN were confirmed to belong to this nucleus. However, 6/34 MERs that were electrophysiologically attributed to the ventral STN were postoperatively reattributed to the SN. Furthermore, 44 MERs of 3 trajectories, which were intraoperatively attributed to the STN, were postoperatively reattributed to the pallidum or thalamus. MER parameters seemed to differ across the STN, with higher spike amplitudes (H = 10.64, p < 0.01) and less prevalent β-oscillations (H = 9.81, p < 0.01) in the limbic STN than in the sensorimotor and associative subdivisions. Some cells, especially in the SN, showed longer spikes with lower firing rates, in agreement with described characteristics of dopamine cells. However, these probabilistic electrophysiological signatures might become clinically less relevant with the development of image reconstruction tools, which deserve to be applied intraoperatively.
... The findings were inferred based on the seed regions extracted from the ascending arousal network. Coordinates of the seed regions may improve as newer standards are created [40]. The actual metric being compared is the 2-back>0-back differential activity during the difficult cognitive working memory task. ...
Preprint
Background: Myalgic Encephalomyelitis / Chronic Fatigue Syndrome (ME/CFS), Gulf War Ill-ness (GWI) and control subjects had fMRI during difficult cognitive tests performed before and after submaximal exercise provocation (Washington 2020). Exercise caused increased activation in ME/CFS but decreased activation for GWI in the dorsal midbrain, left Rolandic operculum and right middle insula. Midbrain and isthmus nuclei participate in threat assessment, attention, cognition, mood, pain, sleep, and autonomic dysfunction Methods: Activated midbrain nuclei were inferred by re-analysis of data from 31 control, 36 ME/CFS and 78 GWI subjects using a seed region approach and the Harvard Ascending Arousal Network. Results: Before exercise, control and GWI had greater activation during cognition than ME/CFS in left pedunculotegmental nucleus. Postexercise ME/CFS had greater activation than GWI for midline periaqueductal gray, dorsal and median raphe, and right midbrain reticular formation, parabrachial complex and locus coeruleus. The change between days (delta) was positive for ME/CFS but negative for GWI indicating reciprocal patterns of activation. Controls had no changes. Conclusions: Exercise caused opposite effects with increased activation in ME/CFS but decreased activation in GWI indicating different pathophysiological responses to exertion and mechanisms of disease. Midbrain and isthmus nuclei contribute to postexertional malaise in ME/CFS and GWI.
Article
Full-text available
Patients with psychiatric symptoms, such as depression, anxiety, and visual hallucinations, may be at increased risk for adverse effects following deep brain stimulation of the subthalamic nucleus for Parkinson’s disease, but there have been relatively few studies of associations between locations of chronic stimulation and neuropsychological outcomes. We sought to determine whether psychiatric history modulates associations between stimulation location within the subthalamic nucleus and postoperative affective and cognitive changes. We retrospectively identified 42 patients with Parkinson’s disease who received bilateral subthalamic nucleus deep brain stimulation and who completed both pre- and postoperative neuropsychological testing. Active stimulation contacts were localized in MNI space using Lead-DBS software. Linear discriminant analysis identified vectors maximizing variance in postoperative neuropsychological changes, and Pearson’s correlations were used to assess for linear relationships. Stimulation location was associated with postoperative change for only 3 of the 18 neuropsychological measures. Variation along the superioinferior (z) axis was most influential. Constraining the analysis to patients with a history of depression revealed 10 measures significantly associated with active contact location, primarily related to location along the anterioposterior (y) axis and with worse outcomes associated with more anterior stimulation. Analysis of patients with a history of anxiety revealed 5 measures with location-associated changes without a predominant axis. History of visual hallucinations was not associated with significant findings. Our results suggest that a history of depression may influence the relationship between active contact location and neuropsychological outcomes following subthalamic nucleus deep brain stimulation. These patients may be more sensitive to off-target (nonmotor) stimulation.
Article
Parkinson’s disease (PD) is treated effectively by deep brain stimulation (DBS) of the subthalamic nucleus (STN), using an electrode inserted into the head of a PD patient. The electrode has multiple electrical contacts along its length, so the best may be chosen for selectively stimulating the STN. Neurosurgeons usually determine the optimal stimulated contact via the clinical experience of the neurosurgeon and the motor improvement of PD patients. This is a time-consuming and labor-intensive trial-and-error process. The selection of optimal stimulated contact highly depends on the locations of sweet spots, which are manually identified by the characteristic features of microelectrode recordings (MERs). This paper presents an amplitude-frequency-aware deep fusion network for optimal contact selection on STN-DBS electrodes. The method first obtains the amplitude-frequency fusion features by combining the MERs time sequence features and the amplitude sequence features, and then uses the convolutional neural network (CNN) with convolutional block attention module (CBAM) to identify both the border of the STN and the sweet spots to implant the electrode. The optimal stimulated contact can be selected according to the distribution of the sweet spots. Experimental results indicate that, for successful surgeries, neurosurgeons and the proposed AI solution selected the same optimal contacts. Furthermore, the proposed method outperforms the state-of-the-art methods for STN and sweet spot identification. The proposed method shows great potential for optimal contact selection to improve the efficiency of STN-DBS surgery and reduce the dependence on clinicians’ experience.
Article
Analysis of the basal ganglia has been important in investigating the effects of Parkinson's disease as well as treatments for Parkinson's disease. One method of analysis has been using MRI for non-invasively segmenting the basal ganglia, then investigating significant parameters that involve the basal ganglia, such as fiber orientations and positional markers for deep brain stimulation. Following enhancements to optimizations and improvements to 3T and 7T MRI acquisitions, we utilized Lead-DBS on human connectome project data to automatically segment the basal ganglia of 49 human connectome project subjects, reducing the reliance on manual segmentation for more consistency. We generated probabilistic tractography streamlines between each segmentation pair using 3T and 7T human connectome diffusion data to observe any major differences in tractography streamline patterns that can arise due to tradeoffs from different field strengths and acquisitions. Tractography streamlines generated between basal ganglia structures using 3T images showed less standard deviation in streamline count than using 7T images. Mean tractography streamline counts generated using 3T diffusion images were all higher in count than streamlines generated using 7T diffusion images. We illustrate a potential method for analyzing the structural connectivity between basal ganglia structures, as well as visualize possible differences in probabilistic tractography that can arise from different acquisition protocols.
Chapter
In this chapter, we will give an overview of important imaging concepts in the field of DBS before going into details in the subsequent chapters. We begin by motivating why imaging in the context of DBS is crucial and which additional scientific questions we can ask if we are able to create meaningful models of DBS. We then discuss several strategies that we can use to increase the precision of our models. The first strategy is to acquire optimal imaging data before DBS surgery. Second, we discuss how normative models of anatomy can be helpful to further augment the definition of the anatomical context surrounding DBS electrodes. Third, we illustrate the importance of methodology to precisely register the former and the latter. We conclude by showing empirical data that justify the use of specialized DBS imaging pipelines over off-the-shelf neuroimaging routines that were originally developed in a different context.
Article
Full-text available
Medical imaging has played a major role in defining the general anatomical targets for deep brain stimulation (DBS) therapies. However, specifics on the underlying brain circuitry that is directly modulated by DBS electric fields remain relatively undefined. Detailed biophysical modeling of DBS provides an approach to quantify the theoretical responses to stimulation at the cellular level, and has established a key role for axonal activation in the therapeutic mechanisms of DBS. Estimates of DBS-induced axonal activation can then be coupled with advances in defining the structural connectome of the human brain to provide insight into the modulated brain circuitry and possible correlations with clinical outcomes. These pathway-activation models (PAMs) represent powerful tools for DBS research, but the theoretical predictions are highly dependent upon the underlying assumptions of the particular modeling strategy used to create the PAM. In general, three types of PAMs are used to estimate activation: 1) field-cable (FC) models, 2) driving force (DF) models, and 3) volume of tissue activated (VTA) models. FC models represent the "gold standard" for analysis but at the cost of extreme technical demands and computational resources. Consequently, DF and VTA PAMs, derived from simplified FC models, are typically used in clinical research studies, but the relative accuracy of these implementations is unknown. Therefore, we performed a head-to-head comparison of the different PAMs, specifically evaluating DBS of three different axonal pathways in the subthalamic region. The DF PAM was markedly more accurate than the VTA PAMs, but none of these simplified models were able to match the results of the patient-specific FC PAM across all pathways and combinations of stimulus parameters. These results highlight the limitations of using simplified predictors to estimate axonal stimulation and emphasize the need for novel algorithms that are both biophysically realistic and computationally simple.
Article
Full-text available
Background: Although recently introduced directional DBS leads provide control of the stimulation field, programing is time-consuming. Objectives: Here, we validate local field potentials recorded from directional contacts as a predictor of the most efficient contacts for stimulation in patients with PD. Methods: Intraoperative local field potentials were recorded from directional contacts in the STN of 12 patients and beta activity compared with the results of the clinical contact review performed after 4 to 7 months. Results: Normalized beta activity was positively correlated with the contact's clinical efficacy. The two contacts with the highest beta activity included the most efficient stimulation contact in up to 92% and that with the widest therapeutic window in 74% of cases. Conclusion: Local field potentials predict the most efficient stimulation contacts and may provide a useful tool to expedite the selection of the optimal contact for directional DBS. © 2017 The Authors. Movement Disorders published by Wiley Periodicals, Inc. on behalf of International Parkinson and Movement Disorder Society.
Article
Full-text available
Deep brain stimulation (DBS) is a neurosurgical intervention where electrodes are permanently implanted into the brain in order to modulate pathologic neural activity. The post-operative reconstruction of the DBS electrodes is important for an efficient stimulation parameter tuning. A major limitation of existing approaches for electrode reconstruction from post-operative imaging that prevents the clinical routine use is that they are manual or semi-automatic, and thus both time-consuming and subjective. Moreover, the existing methods rely on a simplified model of a straight line electrode trajectory, rather than the more realistic curved trajectory. The main contribution of this paper is that for the first time we present a highly accurate and fully automated method for electrode reconstruction that considers curved trajectories. The robustness of our proposed method is demonstrated using a multi-center clinical dataset consisting of N = 44 electrodes. In all cases the electrode trajectories were successfully identified and reconstructed. In addition, the accuracy is demonstrated quantitatively using a high-accuracy phantom with known ground truth. In the phantom experiment, the method could detect individual electrode contacts with high accuracy and the trajectory reconstruction reached an error level below 100 micron (0.046 ± 0.025 mm). An implementation of the method is made publicly available such that it can directly be used by researchers or clinicians. This constitutes an important step towards future integration of lead reconstruction into standard clinical care.
Article
Full-text available
View largeDownload slide The use of subthalamic nucleus deep brain stimulation in Parkinson’s disease is limited in some patients by behavioural side-effects. Using perioperative electrophysiological recordings and presurgical neuroimaging, Accolla et al . characterize the anatomical networks modulated by deep brain stimulation, and reveal the existence of overlapping functional areas within the nucleus. View largeDownload slide The use of subthalamic nucleus deep brain stimulation in Parkinson’s disease is limited in some patients by behavioural side-effects. Using perioperative electrophysiological recordings and presurgical neuroimaging, Accolla et al . characterize the anatomical networks modulated by deep brain stimulation, and reveal the existence of overlapping functional areas within the nucleus.
Article
BACKGROUND: Targeting accuracy in deep brain stimulation (DBS) surgery can be defined as the level of accordance between selected and anatomic real target reflected by characteristic electrophysiological results of microelectrode recording (MER). OBJECTIVE: To determine the correspondence between the preoperative predicted target based on modern 3-T magnetic resonance imaging (MRI) and intraoperative MER results separately on the initial and consecutive second side of surgery. METHODS: Retrospective cohort study of 86 trajectories of DBS electrodes implanted into the subthalamic nucleus (STN) of patients with Parkinson's disease. The entrance point of the electrode into the STN and the length of the electrode trajectory crossing the STN were determined by intraoperative MER findings and 3 T T2-weighted magnetic resonance images with 1-mm slice thickness. RESULTS: Average difference between MRI- and MER-based trajectory lengths crossing the STN was 0.28 ± 1.02 mm (95% CI: −0.51 to −0.05 mm). There was a statistically significant difference between the MRI- and MER-based entry points on the initial and second side of surgery (P = .04). Forty-three percent of the patients had a difference of more than ±1 mm of the MRI-based-predicted and the MER-based-determined entry points into the STN with values ranging from −3.0 to + 4.5 mm. CONCLUSION: STN MRI-based targeting is accurate in the majority of cases on the first and second side of surgery. In 43% of implanted electrodes, we found a relevant deviation of more than 1 mm, supporting the concept of MER as an important tool to guide and optimize targeting and electrode placement.
Article
Objectives: Firstly, to identify subthalamic region stimulation clusters that predict maximum improvement in rigidity, bradykinesia and tremor, or emergence of side-effects; and secondly, to map-out the cortical fingerprint, mediated by the hyperdirect pathways which predict maximum efficacy. Methods: High angular resolution diffusion imaging in twenty patients with advanced Parkinson's disease was acquired prior to bilateral subthalamic nucleus deep brain stimulation. All contacts were screened one-year from surgery for efficacy and side-effects at different amplitudes. Voxel-based statistical analysis of volumes of tissue activated models was used to identify significant treatment clusters. Probabilistic tractography was employed to identify cortical connectivity patterns associated with treatment efficacy. Results: All patients responded well to treatment (46% mean improvement off medication UPDRS-III [p < 0.0001]) without significant adverse events. Cluster corresponding to maximum improvement in tremor was in the posterior, superior and lateral portion of the nucleus. Clusters corresponding to improvement in bradykinesia and rigidity were nearer the superior border in a further medial and posterior location. The rigidity cluster extended beyond the superior border to the area of the zona incerta and Forel-H2 field. When the clusters where averaged, the coordinates of the area with maximum overall efficacy was X = -10(-9.5), Y = -13(-1) and Z = -7(-3) in MNI(AC-PC) space. Cortical connectivity to primary motor area was predictive of higher improvement in tremor; whilst that to supplementary motor area was predictive of improvement in bradykinesia and rigidity; and connectivity to prefrontal cortex was predictive of improvement in rigidity. Interpretation: These findings support the presence of overlapping stimulation sites within the subthalamic nucleus and its superior border, with different cortical connectivity patterns, associated with maximum improvement in tremor, rigidity and bradykinesia.
Article
Three-dimensional atlases of subcortical brain structures are valuable tools to reference anatomy in neuroscience and neurology. For instance, they can be used to study the position and shape of the three most common deep brain stimulation (DBS) targets, the sub-thalamic nucleus (STN), internal part of the pallidum (GPi) and ventral intermediate nucleus of the thalamus (VIM) in stereotactic space and in spatial relationship to DBS electrodes. Here, we present a composite atlas that is based on manual segmentations of a multimodal high resolution brain template, histology and structural connectivity. In a first step, four key structures were defined on the template itself using a combination of multi-spectral image analysis and manual segmentation. Second, these structures were used as anchor points to coregister a detailed histological atlas into standard space. Results show that this approach significantly improved coregistration accuracy over previously published methods. Finally, a sub-segmentation of STN and GPi into functional zones was achieved based on structural connectivity. The result is a composite atlas that defines key nuclei on the template itself, fills the gaps between them using histology and further subdivides them using structural connectivity. We show that the atlas can be used to segment DBS targets in single subjects, yielding more accurate results compared to priorly published atlases. The atlas will be made publicly available and constitutes a resource to study DBS electrode localizations in combination with modern neuroimaging methods.
Article
In neurosurgical literature, findings such as deep brain stimulation (DBS) electrode positions are conventionally reported in relation to the anterior and posterior commissures of the individual patient (AC/PC coordinates). However, the neuroimaging literature including neuroanatomical atlases, activation patterns, and brain connectivity maps has converged on a different population-based standard (MNI coordinates). Ideally, one could relate these two literatures by directly transforming MRIs from neurosurgical patients into MNI space. However obtaining these patient MRIs can prove difficult or impossible, especially for older studies or those with hundreds of patients. Here, we introduce a methodology for mapping an AC/PC coordinate (such as a DBS electrode position) to MNI space without the need for MRI scans from the patients themselves. We validate our approach using a cohort of DBS patients in which MRIs are available, and test whether several variations on our approach provide added benefit. We then use our approach to convert previously reported DBS electrode coordinates from eight different neurological and psychiatric diseases into MNI space. Finally, we demonstrate the value of such a conversion using the DBS target for essential tremor as an example, relating the site of the active DBS contact to different MNI atlases as well as anatomical and functional connectomes in MNI space.
Article
OBJECTIVE Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is a well-established therapy for motor symptoms in patients with pharmacoresistant Parkinson's disease (PD). However, the procedure, which requires multimodal perioperative exploration such as imaging, electrophysiology, or clinical examination during macrostimulation to secure lead positioning, remains challenging because the STN cannot be reliably visualized using the gold standard, T2-weighted imaging (T2WI) at 1.5 T. Thus, there is a need to improve imaging tools to better visualize the STN, optimize DBS lead implantation, and enlarge DBS diffusion. METHODS Gradient-echo sequences such as those used in T2WI suffer from higher distortions at higher magnetic fields than spin-echo sequences. First, a spin-echo 3D SPACE (sampling perfection with application-optimized contrasts using different flip angle evolutions) FLAIR sequence at 3 T was designed, validated histologically in 2 nonhuman primates, and applied to 10 patients with PD; their data were clinically compared in a double-blind manner with those of a control group of 10 other patients with PD in whom STN targeting was performed using T2WI. RESULTS Overlap between the nonhuman primate STNs segmented on 3D-histological and on 3D-SPACE-FLAIR volumes was high for the 3 most anterior quarters (mean [± SD] Dice scores 0.73 ± 0.11, 0.74 ± 0.06, and 0.60 ± 0.09). STN limits determined by the 3D-SPACE-FLAIR sequence were more consistent with electrophysiological edges than those determined by T2WI (0.9 vs 1.4 mm, respectively). The imaging contrast of the STN on the 3D-SPACE-FLAIR sequence was 4 times higher (p < 0.05). Improvement in the Unified Parkinson's Disease Rating Scale Part III score (off medication, on stimulation) 12 months after the operation was higher for patients who underwent 3D-SPACE-FLAIR–guided implantation than for those in whom T2WI was used (62.2% vs 43.6%, respectively; p < 0.05). The total electrical energy delivered decreased by 36.3% with the 3D-SPACE-FLAIR sequence (p < 0.05). CONCLUSIONS 3D-SPACE-FLAIR sequences at 3 T improved STN lead placement under stereotactic conditions, improved the clinical outcome of patients with PD, and increased the benefit/risk ratio of STN-DBS surgery.