AGE-RELATED CHANGES OF VOLUME AND POSITION OF THE SUBTHALAMIC NUCLEUS

AGE-RELATED CHANGES OF VOLUME AND POSITION OF THE

SUBTHALAMIC NUCLEUS

Cognitive Neuroscience track

Supervisor Max Keuken

Universiteit van Amsterdam Amsterdam Brain and Cognition

Co-Assessor

Prof. Dr. Birte Forstmann

Universiteit van Amsterdam Amsterdam Brain and Cognition

     UvA Representative Dr. Leendert van Maanen

Universiteit van Amsterdam Amsterdam Brain and Cognition

Age-related changes of volume and position of the subthalamic nucleus

Abstract

The subthalamic nucleus (STN) is a small subcortical structure, whose involvement in motor functions makes it a prime target for deep brain stimulation (DBS) in Parkinson’s disease. As some studies suggest that the STN might be subject to age related changes while others claim no differences in anatomy in higher age, this study aimed to replicate previous findings on age related changes. Because available standard atlases do not take these changes into account, an atlas of the STN for a young, middle-aged and elderly group was created. Ultra-high field 7 tesla (T) magnetic resonance imaging (MRI) was used to manually segment the STN and to build a probabilistic atlas. Results show a decrease in volume and lateral shift with increasing age. The probabilistic atlas was also used to analyze the DBS target location reported in 24 studies. We found a general good overlap suggesting that DBS stimulation targets the STN accurately according to our atlas. Overall, the present study demonstrates that age acts as an important factor in anatomical changes in the STN and that these changes have to be taken into account in brain atlases to ensure exact localization of the STN across different age groups.

1 Introduction

The subthalamic nucleus (STN) is a small almond-shaped subcortical structure, lying medial to the substantia nigra and lateral to the hypothalamus (den Dunnen & Staal, 2005; Massey & Yousry, 2010; Schaltenbrand & Wahren, 1977). Because it is involved in motor functions (Weintraub & Zaghloul, 2013) the STN is targeted in deep brain stimulation (DBS) in Parkinson’s Disease (PD) to treat motor symptoms in patients that are unresponsive to non-surgical medical treatments (Benabid et al., 2000; Daniluk et al., 2010). Benefits of this stimulation can be a reduction of the severity of symptoms like rigidity, tremors, bradykinesia and dyskenisia even though the underlying mechanisms are not completely known (Benabid et al., 2009).

There are several different techniques to locate the correct position for DBS electrodes (Daniluk et al., 2010). One method that takes the patients’ individual anatomy into account is to directly visualize the STN by using magnetic resonance imaging (MRI). However, the STN is difficult to visualize by standard MRI protocols because of its hypointense signal and because spatial resolution is often not sufficient to show the small structure (Rong et al., 2013). MRI images are usually acquired with in MR scanners with field strength of 1.5 T or 3 T (Cho et al., 2010). Especially in 1.5T MR images the borders of the STN cannot be directly observed as the spatial resolution and contrast do not suffice (Cheng et al., 2013; Cho et al., 2010). MR imaging at a field strength of 7T allows for the exact visualization of the borders of the STN due to its enhanced resolution (Beisteiner et al., 2011; Cho et al., 2010; Tourdias et al., 2014) and can thus also be used to determine the position of structures as a basis for atlases, as those measurements are generally stable (Han et al., 2006) and localization is highly reliable (Buchanan et al., 2014; Jovicich et al., 2013; Wonderlick et al., 2009). However, 7T MRI scanners are not yet commonly available (Cho et al., 2010). For that reason, an alternative method that is often used is indirect targeting of the STN. In this method, coordinates for the position of the STN are calculated relative to the patient’s anterior commissure to posterior commissure line (AC-PC line; (Daniluk et al., 2010). The coordinates are usually derived from the human brain atlas by Schaltenbrand and Wahren (1977). However, comparisons of STN coordinates derived from direct MRI visualization and coordinates derived from a standard atlas have shown that this method can hold inaccuracies (Ashkan et al., 2007; Hariz et al., 2004; Schlaier et al., 2005). A possible reason for these could be that this brain atlas is based on only three human brains (Niemann & van Nieuwenhofen, 1999). Anatomical variability of individuals is thus not depicted adequately in the atlas. Also, age is not taken into account as a factor that can influence structural MRI findings, as global brain changes go along with normal aging (Rodrigue et al., 2011).

Because anatomical changes in the brain can influence the targeting of the STN, previous research has focused on changes of the STN over age. However, there doesn’t seem to be a consensus regarding the findings across different studies, making definite conclusions questionable. Possible changes of the anatomy of the STN that have been reported previously are its position in regard to the AC-PC line and a change in volume. A seemingly common change due to ageing is the change in volume. Grey matter tissue generally is reported to decrease whereas ventricle volumes increase (Evans et al., 2012; Good et al., 2001). Furthermore, shrinkage of several subcortical structures in close proximity to the STN has been reported previously (Cherubini et al., 2009; Fjell et al., 2013; Greenberg et al., 2008; Raz et al., 2010; Raz et al., 2005; Walhovd et al., 2005; Walhovd et al., 2011). It is unlikely that neighboring structures develop completely independently of each other (Raz et al., 2005), so it is to be expected that the STN volume decreases with age. Some studies have shown a

reduction of STN volume over age (den Dunnen & Staal, 2005; Shen et al., 2009) while others found no significant change (Keuken et al., 2013).

These possible volume changes might occur in combination with position changes of the STN. However, different results have been found regarding its position in the ageing brain. While a lateral shift of the STN over age was shown post-mortem (den Dunnen & Staal, 2005) as well as in-vivo (Keuken et al., 2013; Kitajima et al., 2008), other post-mortem findings suggest that the location of the STN is stable over age (Massey et al., 2012). Mavridis (2013) even found an opposite effect, as the STN in young subjects was located more laterally compared to its position in subjects over the age of 60.

This study aims to replicate the shift found by Keuken et al. (2013) in a new cohort of subjects using 7T MRI and to test whether there is a volume shrinkage of the STN over age. By using the same methods as Keuken and colleagues, this would not only show changes of the STN over age, but also stress the importance of replication, give clues about the reliability of 7T imaging and the dependence of brain atlases on interindividual anatomical differences.

The data of this study will be used to create a four dimensional probabilistic atlas of the STN for three age groups that will be made publicly available. In order to visualize whether placement of electrodes in DBS targets the STN as this atlas defines it for a middle aged and elderly population we also analyzed stimulation site coordinates given by 24 studies studies (Ashkan et al., 2007; Chen et al., 2011; Guo et al., 2013; Hamel et al., 2003; Hamid et al., 2005; Herzog et al., 2004; Lalys et al., 2013; Lefranc et al., 2014; Littlechild et al., 2003; Liu et al., 2013; Nestor et al., 2014; Novak et al., 2011; Ostrem et al., 2011; Pollo et al., 2007; Reese et al., 2012; Rocchi et al., 2012; Saint-Cyr et al., 2002; Sarnthein et al., 2013; Smith et al., 2014; Toda et al., 2009; Tsai et al., 2007; Witt et al., 2013; Yokoyama et al., 2006; Zonenshayn et al., 2004) as well as the coordinates suggested by the Schaltenbrand and Wahren atlas (Schaltenbrand & Wahren, 1977). The atlases for the middle-aged and elderly population were chosen for a comparison with these DBS sites because the given studies reported implantations in patients within this age-range.

2 Methods 2.1 Participants

Brain scans of 28 young participants, who were between 19 and 29 years old (mean 24, SD 2.27, 14 male, 14 female) were analyzed in this study. Additionally, ten middle aged subjects between 40 and 60 years (mean 51.89, SD 7.57, 5 male, 4 female) and seven elderly participants between 62 and 74 years (mean 68.0, SD 4.55, 4 male, 3 female) were scanned. All participants had no diagnosis or history of psychiatric or neurological disorders. All subjects were fully informed and gave written consent before MRI images were acquired and were compensated monetarily. The study was approved by the ethics committee at the University of Leipzig, Germany.

2.2 Data Acquisition

All participants were scanned in a 7T Magnetom MRI system (Siemens, Erlangen) using a 24-channel head array Nova Coil (NOVA Medical Inc., Wilmington MA). An MP2RAGE (repetition time (TR) = 5000ms, echo time (TE) = 2.45ms, inversion times TI1/TI2 900/2750ms, flip angle = 5°/3°, bandwidth = 250 Hz/Px) was used to acquire a whole brain image of 240 sagittal slices with 0.7mm isotropic resolution. Additionally, a zoomed MP2RAGE slab with 0.6mm isotropic resolution (TR= 5000ms, TE =

3.71ms, TI1/TI2 = 900/2750ms, flip angle = 5°/3°, bandwidth 240 Hz/Px, 128 slices) and a zoomed multi-echo 3D FLASH slab with 0.5mm isotropic resolution (TR = 41ms, TEs = 11.22/20.39/29.57ms, flip angle = 14°, bandwidth 160 Hz/Px, 128 slices) were collected in axial slices parallel to the AC-PC line. Only the 3D FLASH slab was used for segmentation of the STN. Both MP2RAGE scans were used for registrations of the STN masks to MNI space.

2.3 Manual Segmentation

The STN was segmented manually by two independent researchers in FSL 5.0.2 viewer (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/). Both followed a segmentation protocol that specifies anatomical landmarks to localize and identify the STN. To assess inter-rater agreement, Cohen’s Kappa (Cohen, 1960) was calculated for all three groups (mean/SD of Cohen’s k : young: k = 0.68/0.14, middle-aged: k = 0.71/0.07, elderly: 0.65/0.12). The masks used for analysis of volume and position changes consisted only of the overlapping voxels that both raters had marked as belonging to the STN.

2.4 Registration

Masks were registered using MIPAV 5.4.4. (http://mipav.cit.nih.gov/). All scans were skull-stripped before registration. This was done by the implemented MP2RAGE skull strip algorithm in the CBS High-Res Brain Processing Tools for MIPAV (http://www.cbs.mpg.de/institute/software/obs-hrt/ondex.html). FLASH scans were stripped using BET as implemented in FSL 5.0.2 (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/). For all registration steps an optimized automatic linear registration algorithm with 12 degrees of freedom was used. The FLASH slabs were registered to the MP2RAGE slab. Those were registered to the whole brain MP2RAGE scans, which were finally registered to the 0.4 mm resolution MNI template. The STN masks were consecutively transformed to MNI space using the transformation matrices that were created in the former registration steps. 2.5 Analysis

Mixed effect models were calculated to analyze a change of volume of the STN over age by subsequently adding factors until no better prediction levels in dependence on levels of freedom could be acquired using R (http://www.R-project.com). To maintain the individual anatomical variability this analysis was done in native space before registration to MNI space.

Possible shifts of the STN location over age were analyzed by running regressions on x, y and z coordinates of the centers of mass for all masks in R (http://www.R-project.com). These coordinates were obtained using FSUTILS in FSL 5.0.2. and given in voxel numbers. For all regressions outliers were excluded by calculating Cook’s distance. 4/44 (4/number of observations) was chosen as a cutoff-criterium, which lead to the exclusion of 2 subjects for regressions of x-coordinates and the y-coordinate in the right hemisphere, 3 excluded subjects for regression of z-y-coordinates and the exclusion of 4 subjects for the regression of the y-coordinate in the left hemisphere.

2.6 Building of a probabilistic atlas and comparison with DBS coordinates

For the probabilistic atlas, all masks within age groups were combined using FSLUTILS in FSL 5.0.2. 24 studies on DBS in PD were found by a literature search on Pubmed (http://www.ncbi.nlm.nih.gov/pubmed/) using various keywords (e.g. “DBS”, “Parkinson’s disease”,

“position”, “coordinate”). As most of the stimulation coordinates were given in distance to the midcommissural point, this was calculated for the MNI04 template. AC and PC were manually determined in the FSL viewer resulting in coordinates of x =0.10, y = 0.02, z = 4.13 and x = 0.10, y = -24.78, z = -1.73, respectively. The coordinates of the midcommissural point were thus defined as x = 0.10, y = -12.38 and z = -2.93 and were used to calculate the coordinates used in the comparison with the probabilistic atlas. Then masks for every study were created by constructing a cube of 2 mm edge length around the calculated coordinate. All masks were visualized with the middle-aged and elderly atlas in FSL 5.0.2 and DBS positions were visually evaluated. Both atlases were displayed with a threshold of 20% of their maximal overlap to exclude very low probabilities from the comparison. 3 Results

3.1 Volume changes of the STN over age

The STN volumes of the young, middle-aged and elderly group are displayed in Table 1. We found a significant decline of STN volume over age as shown in Figure 1. A regression with fixed effects for age and Cohen’s kappa for single masks as well as their interaction and random effects for subjects was run on all masks. These factors explained 52.0% percent of the variance in the data (R2 = 0.520, F(3,84) = 15.65, p < 0.001).

< Table 1 around here > < Figure 1 around here > 3.2 Position changes of the STN over age

We found a lateral shift of the x coordinate of the center of mass of the STN in both hemispheres (Figure 2). For the left hemisphere, a regression of age and Cohen’s kappa on the x-coordinate was significant (F(2,39) = 6.479, p < 0.01, r2 = 0.49) as was the regression in the right hemisphere (F(2,39) = 13.77, p < 0.001, r2 = - 0.59). No age effect was found for the y- or z- coordinates of the center of mass (p > 0.05).

< Figure 2 around here > 3.3 Probabilistic atlas and comparison with DBS coordinates

Combining the interrater STN masks resulted in probabilistic atlases for the young, middle-aged and elderly with maximal overlaps of 43.67%, 28.26% and 38.04% respectively. We compared positions reported for DBS sites in 24 studies (Ashkan et al., 2007; Chen et al., 2011; Guo et al., 2013; Hamel et al., 2003; Hamid et al., 2005; Herzog et al., 2004; Lalys et al., 2013; Lefranc et al., 2014; Littlechild et al., 2003; Liu et al., 2013; Nestor et al., 2014; Novak et al., 2011; Ostrem et al., 2011; Pollo et al., 2007; Reese et al., 2012; Rocchi et al., 2012; Saint-Cyr et al., 2002; Sarnthein et al., 2013; Smith et al., 2014; Toda et al., 2009; Tsai et al., 2007; Witt et al., 2013; Yokoyama et al., 2006; Zonenshayn et al., 2004) with the middle-aged and elderly atlas, which were both displayed with a threshold of 20% of their maximal overlap (Figure 3). A good overlap generally was found, with most electrodes placed in the more posterior and superior part of the STN as we mapped it in the middle-aged as well as the elderly subjects. For the middle aged, two reported coordinates (Smith et al., 2014; Tsai et al., 2007) showed no overlap with the atlas in the left hemisphere and stimulation sites reported by six studies (Chen et al., 2011; Littlechild et al., 2003; Novak et al., 2011; Rocchi et al., 2012; Smith et al., 2014;

Tsai et al., 2007) were outside the STN borders in the right hemisphere. The analysis for the elderly atlas lead to the same result in the right hemisphere, while three more DBS sites were identified to have no overlap with the atlas in the left hemisphere (Chen et al., 2011; Littlechild et al., 2003; Rocchi et al., 2012).

< Figure 3 around here >

The mean coordinate of all studies is given in Table 2. This point differs less than 0.5 mm in every dimension from the standard coordinates given by the Schaltenbrand and Wahren atlas (12 mm lateral, 4 mm inferior and 2mm posterior to the midcommissural point result in the coordinates given in Table 2; (Schaltenbrand & Wahren, 1977) ). The mean centers of mass of the STN according to our atlases for the middle-aged and elderly are shown in Table 2.

< Table 2 around here > 4 Discussion

4.1 STN volume decrease over age

The shrinkage of the STN we found over age is in agreement with prior findings by den Dunnen and Staal (2005) and Shen et al. (2009), whereas this was not found by the prior study by Keuken et al. (2013). Keuken et al. (2013) argued, that a possible aging effect might not have been found due to a restricted age range. However, as the range in the current study was younger than in the prior study, this explanation is unlikely. Loss of volume in the STN was expected as shrinkage of neighboring subcortical structures (Fjell et al., 2013; Raz et al., 2010; Raz et al., 2005; Walhovd et al., 2005; Walhovd et al., 2011) as well as shrinkage of the STN over age has been shown before (den Dunnen & Staal, 2005; Shen et al., 2009).

Different underlying mechanisms have been proposed for ageing of the brain and changes in its volume. Two changes going along with age and possibly influencing tissue shrinkage are the change in concentrations of certain metabolic markers and a generally elevated blood pressure (Raz & Rodrigue, 2006; Sachdev, 2004). Homocysteine is an amino acid, that originates from the metabolism of methione and is usually quickly transformed into other amino acids (Schalinske & Smazal, 2012). Its concentrations increase over age, which is supposed to contribute to the neuronal loss seen in higher age. Increased levels of homocysteine can lead to mitochondrial dysfunction and increase the neurotoxic effects of beta-amyloids which subsequently leads to further cell death (Sachdev, 2004). Such increased levels of homocysteine have previously been found to be correlated to reduced grey matter volume (Whalley et al., 2003), increased ventricle size (Sachdev et al., 2002) and a decrease in subcortical volume (Sachdev, 2004).

Another possible cause for the decline in brain volume over age can be changes in blood pressure, which is commonly elevated in the elderly (Raz & Rodrigue, 2006). A negative correlation of blood pressure and total as well as regional brain volume has been shown before (Beauchet et al., 2013). Elevated blood pressure can lead to a change of atherosclerotic characteristics of blood vessels. This makes them prone to minor lesions within the brain which consequently lead to neuronal cell death and brain volume loss. (Beauchet et al., 2013). Furthermore, animal models suggest that an elevated blood pressure can lead to a decrease in cerebral blood flow as well as capillary abnormalities. These changes prompt differences in the blood-brain barrier and subsequently promote neuronal death. (Beauchet et al., 2013).

Changes in metabolic marker levels and elevated blood pressure are known to be effects of normal aging (Raz & Rodrigue, 2006; Sachdev, 2004) and could be possible causes for the decrease in STN volume we found. However, neither metabolic marker levels nor blood pressure were measured in this study which makes it impossible to make actual predications on an interrelation of these factors and our findings. To promote an understanding of volume loss in the STN, future studies should thus also include possible underlying factors in studies investigating structural changes over age.

4.2 Shift of the STN in lateral direction over age

Next to this volumetric effect we also found a shift of the center of the STN in lateral direction with increasing age. This finding replicates the results by Keuken et al. (2013) and has previously been shown by other groups (den Dunnen & Staal, 2005; Kitajima et al., 2008). The lateral shift we observed might be caused by an expansion of the ventricles, which is usually observed in higher age and is thought to be due to white matter atrophy (Damkier et al., 2013). Additionally, the cerebrospinal fluid system in the ventricles has an important function as a brain barrier that is also affected by aging (Shook et al., 2014). Cerebral spinal fluid secretion by the choroid plexus declines over age which combined with enlarged ventricles leads to an increased cerebrospinal fluid drainage resistance. Because compounds in the cerebrospinal fluid can consequently not be removed as well, concentrations for some compounds reach higher levels than in younger age when this drainage is unimpaired (Damkier et al., 2013). Thus, the cerebrospinal fluid system loses some of its filtering function (Shook et al., 2014) which is responsible for more neuronal cell death. This additional atrophy could lead to further enlargement of the ventricular system (Damkier et al., 2013) and it might also be an important factor considering brain volume changes over age.

4.3 Probabilistic atlas and comparison with DBS coordinates

The middle-aged and elderly atlas were used to investigate stimulation sites taken from 24 published studies of DBS in PD patients. Overall, this comparison showed a considerable overlap for the middle-aged group. In the elderly atlas however, 25% of the reported positions lay outside of the borders of the STN. This result might be a consequence of inaccuracies or of a number of other factors. For one reason, the coordinates reported in the studies were not given in MNI space (except for Witt et al. (2013)) while our atlases are. Individual brain sizes could thus influence our results, as we mapped coordinates directly in MNI space without any registration steps. Also, only 5 out of 24 studies reported individual coordinates for each hemisphere (Lalys et al., 2013; Liu et al., 2013; Novak et al., 2011; Smith et al., 2014; Witt et al., 2013), while all others just reported one lateral distance implying symmetrical positions. For our atlases, no symmetry was previously implied and STNs were segmented separately. This could explain why our atlas for the middle-aged had a good overlap with DBS positions in the left hemisphere while six positions lay outside the STN in the right hemisphere. The outliers could also be a result of studies not targeting the body of the STN but deliberately targeting the borders as effective stimulation sites (Guo et al., 2013; Hamel et al., 2003; Herzog et al., 2004; Littlechild et al., 2003; Yokoyama et al., 2006; Zonenshayn et al., 2004). For a last reason, the implantation of electrodes relies most heavily on individual anatomy. Some interindividual variability that is not captured in our probabilistic maps could thus also have led to the aberrations we observed.

The mean coordinates of all reported DBS positions showed only little deviation from the coordinate taken from a standard atlas (Schaltenbrand & Wahren, 1977). A probable reason for this great similarity could be that this standard coordinate was used by most studies as an initial orientation in

identifying the STN. The mean coordinate as shown in Figure 3 lay within the borders of the STN in our atlas in the left hemisphere in both the atlas for the middle-aged and elderly. In the right hemisphere however, this position lay on the posterior border of the STN. Our center of mass in both the middle-aged and elderly group differed more from the mean DBS position than the standard coordinate given by Schaltenbrand and Wahren (1977). This could be another effect of some studies not aiming to stimulate in the center of the STN but on its border or close to it.

4.4 Importance of a probabilistic STN atlas

This study provides an atlas of age-dependent high resolution maps of the STN in standard MNI space for three age groups. Interindiviual variability is taken into account in these maps by following a probabilistic approach. These characteristics hold advantages when compared to standard atlases, that are usually based on very small sample sizes and do not include age dependent changes or individual (Evans et al., 2012). This four dimensional approach which includes time dependent changes (Cabezas et al., 2011) is thus an important step in understanding and studying structural properties of the STN.

We created an age-dependent atlas that will be made publicly available (http://www.nitrc.org/projects/atag/). Even though based on small middle-aged and elderly samples, this atlas can be of assistance in identifying the STN in DBS targeting and thus possibly improve medical procedures. This probabilistic atlas can also be of future value by being implemented in fMRI studies. By correctly identifying functional activations in the STN, questions regarding the mechanisms and functions of the STN and its role in more comprehensive brain networks can be answered. Also, this approach could prove useful in understanding which role the STN plays in PD and the mechanisms by which DBS can improve symptoms in PD which are still not understood completely (Benabid et al., 2009).

Also, by replicating findings by Keuken et al. (2013) about a lateral shift of the STN with age this study demonstrates that 7T imaging is a reliable tool to investigate properties of small subcortical structures. Finding a volumetric change that Keuken et al. (2013) did not find using the same methods indicates that the STN might be subject to high individual difference which should be considered in probabilistic atlases.

4.5 Limitations

The main limitation of the current study was that a cross sectional design was used, which needs to be considered when interpreting our results. Cross sectional designs cannot find actual time-dependent changes like measures in longitudinal studies but can only describe differences between groups (Jernigan et al., 2001; Raz & Rodrigue, 2006). These differences can be influenced by other factors than age, because different age groups grew up in different environments and are exposed to different surroundings (Jernigan et al., 2001; Raz & Rodrigue, 2006). However, studies on structural brain changes have shown, that cross sectional results are congruent with longitudinal findings (Fjell et al., 2013) and that cross sectional studies rather underestimate changes over age than overestimating them (Sigurdsson et al., 2012).

The sample sizes in our middle-aged and elderly group with 7 and 9 participants, respectively, are small, considering that probabilistic atlases try to include interindividual variability. However, building age dependent atlases for subcortical structures is at its beginning, and atlases like that by

Keuken et al. (2013) and the atlas created in this study should be viewed as a basis which more extensive projects can build on.

A further limitation of this study was some very low interrater validity values, as two masks had Cohen’s kappas below 0.3. These had no effect on our results however, as both masks were excluded as outliers in the regression of x coordinates. We also repeated the volumetric analysis after excluding both masks, which again did not change our results.

5 Conclusions

This study shows that the STN is subject to age related changes as they have been shown before by Keuken et al. (2013). We found that STN volume decreases and that the STN shifts lateral direction over age. Both these changes should be considered, when working with standard atlases and when targeting the STN in DBS to ensure reliable and accurate targeting. The underlying mechanisms however, are still unclear and should be addressed in future studies. DBS locations in dependence to our atlas were investigated additionally and a good overlap was found. The four-dimensional atlas built in this study will be made publicly available and can contribute to progress in medical procedures for DBS as well as to research of healthy STN function and its role in PD.

Figures and Tables

Table 1: Mean (SD) STN volumes in cubic millimeters for the young, middle-aged and elderly group.

Left right

Young 40.24mm3 (12.32) 51.29mm3 (14.16)

Middle Aged 26.85mm3 (10.02) 26.03mm3 (10.10)

Elderly 30.66mm3 (8.66) 34.27mm3 (18.31)

Table 2: DBS stimulation sites and center of mass. This table shows the BDS site suggested by (Schaltenbrand & Wahren, 1977), the mean stimulation coordinates of 24 analyzed studies and the center of mass for both the atlas built for a middle-aged and an elderly population in this study.

left hemisphere right hemisphere

x y z x y z Schaltenbrand & Wahren , 1977 -11.9 mm -14.38 mm -6.93 mm 12.1 mm -14.38 mm -6.93 mm DBS sites -11.54 mm -14.48 mm -6.54 mm 11.79 mm -14.46 mm -6.58 mm Middle Aged -10.30 mm -12.38 mm -7.33 mm 11.30 mm -11.58 mm -7.33 mm Elderly -11.50 mm -12.78 mm -6.53 mm 11.70 mm -11.98 mm -6.93 mm

Figure 1: Decline of STN volume over age. STN volume is given in cubic millimeters, age is given in years. STN volumes were calculated for single masks in each hemisphere and used to calculate the effect of age in a linear regression.

Figure 2: Lateral shift of the STN over age. 2a) x coordinates are given in MNI space and thus numbers increases from the right to the left side. For the left hemisphere, the center of the STN shifts to the left, while the center of mass shifts to the right in the right hemisphere. 2b) The upper panel of the figure shows a coronal section, the lower panel shows an axial slice in the MNI04 template. The atlas for young subjects is displayed in blue, the atlases for middle-aged and elderly subjects are displayed in green and yellow, respectively. All atlases were visualized with a threshold of 20% of their maximal overlap.

Figure 3: DBS sites and probabilistic atlas. The left panels show the atlas for middle-aged subjects in green, DBS sites in blue and the mean DBS position in yellow. The right panels show the atlas for elderly subjects in red, DBS sites in blue and the mean DBS position in yellow. The top row shows coronal slices, the bottom row shows axial slices in the MNI04 template. Both atlases were displayed with a 20% threshold of their maximal overlap.

Ashkan, K., Blomstedt, P., Zrinzo, L., Tisch, S., Yousry, T., Limousin-Dowsey, P., & Hariz, M. I. (2007). Variability of the subthalamic nucleus: the case for direct MRI guided targeting. Br J

Neurosurg, 21(2), 197-200.

Beauchet, O., Celle, S., Roche, F., Bartha, R., Montero-Odasso, M., Allali, G., & Annweiler, C. (2013). Blood pressure levels and brain volume reduction: a systematic review and meta-analysis. J

Hypertens, 31(8), 1502-1516.

Beisteiner, R., Robinson, S., Wurnig, M., Hilbert, M., Merksa, K., Rath, J., . . . Geissler, A. (2011). Clinical fMRI: evidence for a 7T benefit over 3T. Neuroimage, 57(3), 1015-1021.

Benabid, A. L., Chabardes, S., Mitrofanis, J., & Pollak, P. (2009). Deep brain stimulation of the subthalamic nucleus for the treatment of Parkinson's disease. Lancet Neurol, 8(1), 67-81. Benabid, A. L., Koudsie, A., Benazzouz, A., Fraix, V., Ashraf, A., Le Bas, J. F., . . . Pollak, P. (2000).

Subthalamic stimulation for Parkinson's disease. Arch Med Res, 31(3), 282-289.

Buchanan, C. R., Pernet, C. R., Gorgolewski, K. J., Storkey, A. J., & Bastin, M. E. (2014). Test-retest reliability of structural brain networks from diffusion MRI. Neuroimage, 86, 231-243.

Cabezas, M., Oliver, A., Llado, X., Freixenet, J., & Cuadra, M. B. (2011). A review of atlas-based segmentation for magnetic resonance brain images. Comput Methods Programs Biomed,

104(3), e158-177.

Chen, S. Y., Tsai, S. T., Hung, H. Y., Lin, S. H., Pan, Y. H., & Lin, S. Z. (2011). Targeting the subthalamic nucleus for deep brain stimulation--a comparative study between magnetic resonance images alone and fusion with computed tomographic images. World Neurosurg, 75(1), 132-137; discussion 122-134, 129-131.

Cheng, C. H., Huang, H. M., Lin, H. L., & Chiou, S. M. (2013). 1.5T versus 3T MRI for targeting subthalamic nucleus for deep brain stimulation. Br J Neurosurg.

Cherubini, A., Peran, P., Caltagirone, C., Sabatini, U., & Spalletta, G. (2009). Aging of subcortical nuclei: microstructural, mineralization and atrophy modifications measured in vivo using MRI. Neuroimage, 48(1), 29-36.

Cho, Z. H., Min, H. K., Oh, S. H., Han, J. Y., Park, C. W., Chi, J. G., . . . Lee, K. H. (2010). Direct visualization of deep brain stimulation targets in Parkinson disease with the use of 7-tesla magnetic resonance imaging. J Neurosurg, 113(3), 639-647.

Cohen, J. (1960). A Coefficient of Agreement for Nominal Scales. Educational and Psychological

Measurement, 20(1), 37-46.

Damkier, H. H., Brown, P. D., & Praetorius, J. (2013). Cerebrospinal fluid secretion by the choroid plexus. Physiol Rev, 93(4), 1847-1892.

Daniluk, S., K, G. D., Ellias, S. A., Novak, P., & Nazzaro, J. M. (2010). Assessment of the variability in the anatomical position and size of the subthalamic nucleus among patients with advanced Parkinson's disease using magnetic resonance imaging. Acta Neurochir (Wien), 152(2), 201-210; discussion 210.

den Dunnen, W. F., & Staal, M. J. (2005). Anatomical alterations of the subthalamic nucleus in relation to age: a postmortem study. Mov Disord, 20(7), 893-898.

Evans, A. C., Janke, A. L., Collins, D. L., & Baillet, S. (2012). Brain templates and atlases. Neuroimage,

Fjell, A. M., Westlye, L. T., Grydeland, H., Amlien, I., Espeseth, T., Reinvang, I., . . . Alzheimer Disease Neuroimaging, I. (2013). Critical ages in the life course of the adult brain: nonlinear subcortical aging. Neurobiol Aging, 34(10), 2239-2247.

Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N., Friston, K. J., & Frackowiak, R. S. (2001). A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage,

14(1 Pt 1), 21-36.

Greenberg, D. L., Messer, D. F., Payne, M. E., Macfall, J. R., Provenzale, J. M., Steffens, D. C., & Krishnan, R. R. (2008). Aging, gender, and the elderly adult brain: an examination of analytical strategies. Neurobiol Aging, 29(2), 290-302.

Guo, S., Zhuang, P., Hallett, M., Zheng, Z., Zhang, Y., Li, J., & Li, Y. (2013). Subthalamic deep brain stimulation for Parkinson's disease: correlation between locations of oscillatory activity and optimal site of stimulation. Parkinsonism Relat Disord, 19(1), 109-114.

Hamel, W., Fietzek, U., Morsnowski, A., Schrader, B., Herzog, J., Weinert, D., . . . Mehdorn, H. M. (2003). Deep brain stimulation of the subthalamic nucleus in Parkinson's disease: evaluation of active electrode contacts. J Neurol Neurosurg Psychiatry, 74(8), 1036-1046.

Hamid, N. A., Mitchell, R. D., Mocroft, P., Westby, G. W., Milner, J., & Pall, H. (2005). Targeting the subthalamic nucleus for deep brain stimulation: technical approach and fusion of pre- and postoperative MR images to define accuracy of lead placement. J Neurol Neurosurg

Psychiatry, 76(3), 409-414.

Han, X., Jovicich, J., Salat, D., van der Kouwe, A., Quinn, B., Czanner, S., . . . Fischl, B. (2006). Reliability of MRI-derived measurements of human cerebral cortical thickness: the effects of field strength, scanner upgrade and manufacturer. Neuroimage, 32(1), 180-194.

Hariz, M., Blomstedt, P., & Limousin, P. (2004). The myth of microelectrode recording in ensuring a precise location of the DBS electrode within the sensorimotor part of the subthalamic nucleus. Mov Disord, 19(7), 863-864.

Herzog, J., Fietzek, U., Hamel, W., Morsnowski, A., Steigerwald, F., Schrader, B., . . . Volkmann, J. (2004). Most effective stimulation site in subthalamic deep brain stimulation for Parkinson's disease. Mov Disord, 19(9), 1050-1054.

Jernigan, T. L., Archibald, S. L., Fennema-Notestine, C., Gamst, A. C., Stout, J. C., Bonner, J., & Hesselink, J. R. (2001). Effects of age on tissues and regions of the cerebrum and cerebellum.

Neurobiol Aging, 22(4), 581-594.

Jovicich, J., Marizzoni, M., Sala-Llonch, R., Bosch, B., Bartres-Faz, D., Arnold, J., . . . PharmaCog, C. (2013). Brain morphometry reproducibility in multi-center 3T MRI studies: a comparison of cross-sectional and longitudinal segmentations. Neuroimage, 83, 472-484.

Keuken, M. C., Bazin, P. L., Schafer, A., Neumann, J., Turner, R., & Forstmann, B. U. (2013). Ultra-high 7T MRI of structural age-related changes of the subthalamic nucleus. J Neurosci, 33(11), 4896-4900.

Kitajima, M., Korogi, Y., Kakeda, S., Moriya, J., Ohnari, N., Sato, T., . . . Yamashita, Y. (2008). Human subthalamic nucleus: evaluation with high-resolution MR imaging at 3.0 T. Neuroradiology,

50(8), 675-681.

Lalys, F., Haegelen, C., Mehri, M., Drapier, S., Verin, M., & Jannin, P. (2013). Anatomo-clinical atlases correlate clinical data and electrode contact coordinates: application to subthalamic deep brain stimulation. J Neurosci Methods, 212(2), 297-307.

Lefranc, M., Derrey, S., Merle, P., Tir, M., Constans, J. M., Montpellier, D., . . . Krystkowiak, P. (2014). High-resolution Three-dimensional T2 Star Weighted Angiography (HR 3-D SWAN): An Optimized 3T MRI Sequence for Targeting the Subthalamic Nucleus. Neurosurgery.

Littlechild, P., Varma, T. R., Eldridge, P. R., Fox, S., Forster, A., Fletcher, N., . . . Flintham, S. (2003). Variability in position of the subthalamic nucleus targeted by magnetic resonance imaging and microelectrode recordings as compared to atlas co-ordinates. Stereotact Funct

Neurosurg, 80(1-4), 82-87.

Liu, K. D., Shan, D. E., Kuo, T. B., & Yang, C. C. (2013). The effects of bilateral stimulation of the subthalamic nucleus on heart rate variability in patients with Parkinson's disease. J Neurol,

260(7), 1714-1723.

Massey, L. A., Miranda, M. A., Zrinzo, L., Al-Helli, O., Parkes, H. G., Thornton, J. S., . . . Yousry, T. A. (2012). High resolution MR anatomy of the subthalamic nucleus: imaging at 9.4 T with histological validation. Neuroimage, 59(3), 2035-2044.

Massey, L. A., & Yousry, T. A. (2010). Anatomy of the substantia nigra and subthalamic nucleus on MR imaging. Neuroimaging Clin N Am, 20(1), 7-27.

Mavridis, I. N. (2013). Stereotactic neurosurgical anatomy of the nucleus accumbens: four-year outcomes. Surg Radiol Anat, 35(7), 637-638.

Nestor, K. A., Jones, J. D., Butson, C. R., Morishita, T., Jacobson, C. E. t., Peace, D. A., . . . Okun, M. S. (2014). Coordinate-based lead location does not predict Parkinson's disease deep brain stimulation outcome. PLoS One, 9(4), e93524.

Niemann, K., & van Nieuwenhofen, I. (1999). One atlas - three anatomies: relationships of the Schaltenbrand and Wahren microscopic data. Acta Neurochir (Wien), 141(10), 1025-1038. Novak, P., Przybyszewski, A. W., Barborica, A., Ravin, P., Margolin, L., & Pilitsis, J. G. (2011).

Localization of the subthalamic nucleus in Parkinson disease using multiunit activity. J Neurol

Sci, 310(1-2), 44-49.

Ostrem, J. L., Racine, C. A., Glass, G. A., Grace, J. K., Volz, M. M., Heath, S. L., & Starr, P. A. (2011). Subthalamic nucleus deep brain stimulation in primary cervical dystonia. Neurology, 76(10), 870-878.

Pollo, C., Vingerhoets, F., Pralong, E., Ghika, J., Maeder, P., Meuli, R., . . . Villemure, J. G. (2007). Localization of electrodes in the subthalamic nucleus on magnetic resonance imaging. J

Neurosurg, 106(1), 36-44.

Raz, N., Ghisletta, P., Rodrigue, K. M., Kennedy, K. M., & Lindenberger, U. (2010). Trajectories of brain aging in middle-aged and older adults: regional and individual differences. Neuroimage,

51(2), 501-511.

Raz, N., Lindenberger, U., Rodrigue, K. M., Kennedy, K. M., Head, D., Williamson, A., . . . Acker, J. D. (2005). Regional brain changes in aging healthy adults: general trends, individual differences and modifiers. Cereb Cortex, 15(11), 1676-1689.

Raz, N., & Rodrigue, K. M. (2006). Differential aging of the brain: patterns, cognitive correlates and modifiers. Neurosci Biobehav Rev, 30(6), 730-748.

Reese, R., Pinsker, M. O., Herzog, J., Wodarg, F., Steigerwald, F., Potter-Nerger, M., . . . Volkmann, J. (2012). The atypical subthalamic nucleus--an anatomical variant relevant for stereotactic targeting. Mov Disord, 27(4), 544-548.

Rocchi, L., Carlson-Kuhta, P., Chiari, L., Burchiel, K. J., Hogarth, P., & Horak, F. B. (2012). Effects of deep brain stimulation in the subthalamic nucleus or globus pallidus internus on step initiation in Parkinson disease: laboratory investigation. J Neurosurg, 117(6), 1141-1149. Rodrigue, K. M., Haacke, E. M., & Raz, N. (2011). Differential effects of age and history of

hypertension on regional brain volumes and iron. Neuroimage, 54(2), 750-759.

Rong, J., Wang, Q., Liu, K., Tan, L., Ran, X., Zhang, S., . . . Han, Y. (2013). A new atlas localization approach for subthalamic nucleus utilizing Chinese visible human head datasets. PLoS One,

8(2), e57264.

Sachdev, P. (2004). Homocysteine, cerebrovascular disease and brain atrophy. J Neurol Sci, 226(1-2), 25-29.

Sachdev, P. S., Valenzuela, M., Wang, X. L., Looi, J. C., & Brodaty, H. (2002). Relationship between plasma homocysteine levels and brain atrophy in healthy elderly individuals. Neurology,

58(10), 1539-1541.

Saint-Cyr, J. A., Hoque, T., Pereira, L. C., Dostrovsky, J. O., Hutchison, W. D., Mikulis, D. J., . . . Lozano, A. M. (2002). Localization of clinically effective stimulating electrodes in the human subthalamic nucleus on magnetic resonance imaging. J Neurosurg, 97(5), 1152-1166.

Sarnthein, J., Peus, D., Baumann-Vogel, H., Baumann, C. R., & Surucu, O. (2013). Stimulation sites in the subthalamic nucleus projected onto a mean 3-D atlas of the thalamus and basal ganglia.

Acta Neurochir (Wien), 155(9), 1655-1660.

Schalinske, K. L., & Smazal, A. L. (2012). Homocysteine imbalance: a pathological metabolic marker.

Adv Nutr, 3(6), 755-762.

Schaltenbrand, G., & Wahren, W. (1977). Introduction to stereotaxis with an atlas of the human

brain. Stuttgart: Thieme.

Schlaier, J., Schoedel, P., Lange, M., Winkler, J., Warnat, J., Dorenbeck, U., & Brawanski, A. (2005). Reliability of atlas-derived coordinates in deep brain stimulation. Acta Neurochir (Wien),

147(11), 1175-1180; discussion 1180.

Shen, W. G., Wang, H. Y., Lin, Z. G., Shen, H., Chen, X. G., Fu, Y. L., & Gao, W. P. (2009). Stereotactic localization and visualization of the subthalamic nucleus. Chin Med J (Engl), 122(20), 2438-2443.

Shook, B. A., Lennington, J. B., Acabchuk, R. L., Halling, M., Sun, Y., Peters, J., . . . Conover, J. C. (2014). Ventriculomegaly associated with ependymal gliosis and declines in barrier integrity in the aging human and mouse brain. Aging Cell, 13(2), 340-350.

Sigurdsson, S., Aspelund, T., Forsberg, L., Fredriksson, J., Kjartansson, O., Oskarsdottir, B., . . . Gudnason, V. (2012). Brain tissue volumes in the general population of the elderly: the AGES-Reykjavik study. Neuroimage, 59(4), 3862-3870.

Smith, K. M., O'Connor, M., Papavassiliou, E., Tarsy, D., & Shih, L. C. (2014). Phonemic verbal fluency decline after subthalamic nucleus deep brain stimulation does not depend on number of microelectrode recordings or lead tip placement. Parkinsonism Relat Disord.

Toda, H., Sawamoto, N., Hanakawa, T., Saiki, H., Matsumoto, S., Okumura, R., . . . Hashimoto, N. (2009). A novel composite targeting method using high-field magnetic resonance imaging for subthalamic nucleus deep brain stimulation. J Neurosurg, 111(4), 737-745.

Tourdias, T., Saranathan, M., Levesque, I. R., Su, J., & Rutt, B. K. (2014). Visualization of intra-thalamic nuclei with optimized white-matter-nulled MPRAGE at 7T. Neuroimage, 84, 534-545.

Tsai, S. T., Lin, S. H., Lin, S. Z., Chen, J. Y., Lee, C. W., & Chen, S. Y. (2007). Neuropsychological effects after chronic subthalamic stimulation and the topography of the nucleus in Parkinson's disease. Neurosurgery, 61(5), E1024-1029; discussion E1029-1030.

Walhovd, K. B., Fjell, A. M., Reinvang, I., Lundervold, A., Dale, A. M., Eilertsen, D. E., . . . Fischl, B. (2005). Effects of age on volumes of cortex, white matter and subcortical structures.

Neurobiol Aging, 26(9), 1261-1270; discussion 1275-1268.

Walhovd, K. B., Westlye, L. T., Amlien, I., Espeseth, T., Reinvang, I., Raz, N., . . . Fjell, A. M. (2011). Consistent neuroanatomical age-related volume differences across multiple samples.

Neurobiol Aging, 32(5), 916-932.

Weintraub, D. B., & Zaghloul, K. A. (2013). The role of the subthalamic nucleus in cognition. Rev

Neurosci, 24(2), 125-138.

Whalley, L. J., Staff, R. T., Murray, A. D., Duthie, S. J., Collins, A. R., Lemmon, H. A., . . . Deary, I. J. (2003). Plasma vitamin C, cholesterol and homocysteine are associated with grey matter volume determined by MRI in non-demented old people. Neurosci Lett, 341(3), 173-176. Witt, K., Granert, O., Daniels, C., Volkmann, J., Falk, D., van Eimeren, T., & Deuschl, G. (2013).

Relation of lead trajectory and electrode position to neuropsychological outcomes of subthalamic neurostimulation in Parkinson's disease: results from a randomized trial. Brain,

136(Pt 7), 2109-2119.

Wonderlick, J. S., Ziegler, D. A., Hosseini-Varnamkhasti, P., Locascio, J. J., Bakkour, A., van der Kouwe, A., . . . Dickerson, B. C. (2009). Reliability of MRI-derived cortical and subcortical morphometric measures: effects of pulse sequence, voxel geometry, and parallel imaging.

Neuroimage, 44(4), 1324-1333.

Yokoyama, T., Ando, N., Sugiyama, K., Akamine, S., & Namba, H. (2006). Relationship of stimulation site location within the subthalamic nucleus region to clinical effects on parkinsonian symptoms. Stereotact Funct Neurosurg, 84(4), 170-175.

Zonenshayn, M., Sterio, D., Kelly, P. J., Rezai, A. R., & Beric, A. (2004). Location of the active contact within the subthalamic nucleus (STN) in the treatment of idiopathic Parkinson's disease. Surg