Comparison Global Brain Volume Ratios on Alzheimer’s Disease Using 3D T1 Weighted MR Images

Full text

Avrupa Bilim ve Teknoloji Dergisi
Sayı 18, S. 599-606, Mart-Nisan 2020
© Telif hakkı EJOSAT’a aittir
Araştırma Makalesi
www.ejosat.com ISSN:2148-2683
European Journal of Science and Technology
No. 18, pp. 599-606, March-April 2020
Copyright © 2020 EJOSAT
Research Article
http://dergipark.gov.tr/ejosat 599
Comparison Global Brain Volume Ratios on Alzheimer’s Disease
Using 3D T1 Weighted MR Images
Muhammet Üsame Öziç1
*
, Seral Özşen2
1 Necmettin Erbakan University, Faculty of Engineering and Architecture, Department of Biomedical Engineering, Konya, Turkey (ORCID: 0000-0002-3037-2687)
2 Konya Technical University, Faculty of Engineering and Natural Science, Department of Electrical-Electronics Enginnering, Konya, Turkey (ORCID: 0000-0001-
5332-8665)
(İlk Geliş Tarihi 3 Şubat 2020 ve Kabul Tarihi 17 Mart 2020)
(DOI: 10.31590/ejosat.697446)
ATIF/REFERENCE: Öziç, M. Ü. & Özşen, S. (2020). Comparison Global Brain Volume Ratios on Alzheimer’s Disease Using 3D
T1 Weighted MR Images. Avrupa Bilim ve Teknoloji Dergisi, (18), 599-606.
Abstract
Alzheimer's Disease is a cause of dementia that starts with the loss of cognitive functions. The degeneration that starts in memory-
related areas in the brain spreads to other regions as the disease progresses. Volumetric losses occurring in the brain can be monitored
with high resolution 3D T1-weighted magnetic resonance images. The interpretation of these images is carried out by radiologists in
hospitals. However, since the voxel intensity transitions of the brain regions are not clear in magnetic resonance images, computer-
aided numerical methods are needed. These methods can perform pre-processing, post-processing, segmentation and volume
calculation on magnetic resonance images. In this study, gray matter, white matter, cerebrospinal fluid, total intracranial volume,
parenchyma, and lateral ventricle global volumes were calculated for 70 Alzheimer Patients and 70 Normal Control 3D T1-weighted
magnetic resonance images taken from Open Access Series of Imaging Studies database. SPM8 and MRIcro programs, ALVIN and
VBM8 libraries were used. Since the numerical methods used are found in different programs and libraries, a model is proposed
which combinations should be used. Volumetric results are relative due to the different head sizes in each person. Therefore, the
problem of relativity should be eliminated by proportioning each volume value with another volume value. Twenty different metrics
of the brain were obtained by summing and dividing the six global volume regions obtained in different combinations. Using these
values, it was determined whether there was a statistically significant difference between two groups by independent samples t-test.
The performance of the numerical methods and the statistical results of twenty metrics obtained from global brain volumes were
discussed. After measurements and evaluations, it was observed that the ratio of cerebrospinal fluid volume to gray matter volume
was an important marker in the differential diagnosis of the disease.
Keywords: Alzheimer, Volume, Ratios, SPM8, VBM8, MRI
3B T1 Ağırlıklı MR Görüntüleri Kullanarak Alzheimer Hastalığına
İlişkin Global Beyin Hacim Oranlarının Karşılaştırılması
Öz
Alzheimer Hastalığı bilişsel fonksiyonların kaybı ile başlayan bir demans nedenidir. Beyinde hafıza ile ilgili bölgelerde başlayan
dejenerasyon hastalık ilerledikçe diğer bölgelere yayılmaktadır. Beyinde meydana gelen hacimsel kayıplar yüksek çözünürlüklü 3B
T1 ağırlıklı manyetik rezonans görüntüleri ile izlenebilmektedir. Bu görüntülerin yorumlanması hastanelerde radyologlar tarafından
gerçekleştirilmektedir. Ancak manyetik rezonans görüntülerinde beyin bölgelerinin voksel intensite geçişleri net olmadığından
bilgisayar destekli sayısal yöntemlere ihtiyaç duyulmaktadır. Bu yöntemler manyetik rezonans görüntüleri üzerinde önişleme, s on
işleme, segmentasyon ve hacim hesaplama yapabilmektedir. Bu çalışmada Open Access Series of Imaging Studies veri tabanından
alınan 70 Alzheimer Hasta 70 Normal Kontrol 3B T1 ağırlıklı manyetik rezonans görüntüleri üzerinde bilgisayar destekli sayıs al
yöntemler kullanılarak gri madde, beyaz madde, beyin omurilik sıvısı, total beyin hacmi, parankima ve lateral ventrikül bölgelerinin
hacimleri hesaplanmıştır. Çalışmada SPM8 ve MRIcro programları , ALVIN ve VBM8 kütüphaneleri kullanılmıştır. Kullanılan sayısal
yöntemler farklı program ve kütüphaneler de bulundukları için hangi kombinasyonda kullanılmaları gerektiğini gösteren bir model
*
Sorumlu Yazar: Necmettin Erbakan University, Faculty of Engineering and Architecture, Department of Biomedical Engineering, Konya, Turkey,
ORCID: 0000-0002-3037-2687, muozic@gmail.com

Avrupa Bilim ve Teknoloji Dergisi
e-ISSN: 2148-2683 600
önerilmiştir. Her insanda kafa büyüklüğünün farklı olmasından dolayı hacimsel sonuçlar göreceli olmaktadır. Bundan dolayı her bir
hacim değeri başka bir hacim değeri ile oranlanarak görecelik problemi ortadan kaldırılmalıdır. Elde edilen altı global hacim
bölgesinin farklı kombinasyonlarda toplanması ve bölünmesi ile beyne ait yirmi farklı metrik elde edilmiştir. Bu değerler kullanılarak
bağımsız örneklem t-testi ile iki grup arasında istatistiksel olarak anlamlı bir farklılık olup olmadığı belirlenmiştir. Kullanılan sayısal
yöntemlerin performansı ve global beyin hacimlerinden elde edilen yirmi metriğin istatistiksel sonuçları tartışılmıştır. Ölçümler ve
değerlendirmelerden sonra beyin omurilik sıvısı hacminin gri madde hacmine oranının hastalığın ayırıcı tanısında önemli bir i şaretçi
olduğu gözlemlenmiştir.
Anahtar Kelimeler: Alzheimer, Hacim, Oran, SPM8, VBM8, MRI
1. Introduction
Alzheimer's Disease (AD) is a neurological disorder that starts with aging and forgetfulness. It is not known why the disease
started and there is no cure to stop it. If it can be diagnosed at an early stage, there are some treatments that extend the patient's AD
stage for some time (Dubois et al., 2016). Approximately 50-70% of the causes of dementia seen in the clinic are AD (Selekler, 2010).
While the proportion of many common diseases is decreasing in the world, AD is increasing proportionally (Association, 2019).
According to the Alzheimer's Association of Turkey, 600 thousand families in our country are struggling with AD (Derneği, 2020).
AD begins with simple forgetfulness and volumetric losses occur in memory-related areas of the brain. These volume changes of the
brain regions and their ratios to each other give information about the disease as a biomarker (Holland et al., 2009; Petropoulos,
Sibbitt Jr, & Brooks, 1999; Schuff et al., 2009; Villarreal et al., 2002). These biomarkers are extremely important in terms of diagnosis
of the disease, early phase identification, long-term follow-up (G. B. Frisoni, Fox, Jack, Scheltens, & Thompson, 2010; Vemuri &
Jack, 2010). 3D Magnetic Resonance (MR) imaging is a high-resolution medical imaging technique that shows rigid changes in the
brain (G. B. Frisoni et al., 2010; Vemuri & Jack, 2010). MR images consist of successive slices. Therefore, if a slice is being
evaluated, the previous and next slice of that slice should also be evaluated together. 3D MR images are usually interpreted manually
by radiologists. This interpretation gives specific information about the disease such as volumetric measurements, perimeters,
segmentation, and the amount of atrophy. However, the results of the calculation of these variables may vary from experiment to
experiment and give relative results. It is also time-consuming and capable of making mistakes (Keller & Roberts, 2009). Since the
voxel intensity transition between brain regions is not very clear in MR images, computer-aided digital systems that measure and
segment regions automatically instead of manual interpretations could be an effective solution. With the increasing of numerical
methods, automatic brain analysis programs are continuously being developed. Linux based Freesurfer (Fischl, 2012) and FSL
(Jenkinson, Beckmann, Behrens, Woolrich, & Smith, 2012), MATLAB based Statistical Parametric Mapping (SPM) (Ashburner et al.,
2008) are frequently used in the literature. These programs can perform pre-processing, post-processing and segmentation operations
on images using powerful algorithms.
The brain consists of three basic areas: gray matter (GM), white matter (WM), cerebrospinal fluid (CSF). GM is a region where
functional centers are located. WM is the region where neurons are located under GM. The CSF is a fluid that protects the brain from
outside interventions. Since AD starts with forgetfulness, volumetric losses occur primarily in the memory-related regions. Atrophy
occurs especially in the hippocampus, amygdala, limbic system and temporal lobe areas on GM (Petrella, Coleman, & Doraiswamy,
2003). It is stated that there are also some losses in the WM (Salat et al., 2009). As volumetric losses in GM and WM regions increase,
the amount of CSF increases. The lateral ventricle (LV) in the middle of the brain enlarges, and this part is filled with CSF. As a result,
GM, WM, CSF and LV volumes are the main global volume biomarkers for the disease. In this study, GM, WM, CSF and LV volume
measurements were calculated by using packet programs on 3D T1-weighted Normal Control (NC) and AD MR images. 70 AD and
70 NC 3D T1-weighted MR images were taken from the Open Access Series of Imaging Studies (OASIS) database. MR images have
been labeled with Clinical Dementia Rating (CDR) neuropsychological test. Using this test, some questions are asked to patients and
their relatives. Thus, the stage of the disease is determined. CDR consists of five phases; CDR0 normal, CDR0.5 mild cognitive
impairment, CDR1 early stage, CDR2 moderate stage, CDR3 heavy stage. There are data labeled between CDR0-CDR2 in the OASIS
database. In Figure 1, a single slice of MR images labeled CDR0, CDR0.5, CDR1, CDR2 are given, respectively. As seen in Figure 1,
as the stage progresses, the brain shrinks and volumetric losses increase (Öziç, 2018).
Figure 1: Single slice of MR images labeled CDR0, CDR0.5, CDR1, CDR2, respectively (Marcus et al., 2007)
MRIcro, SPM8, VBM8, ALVIN package programs were used in this study (Kempton et al., 2011; Kurth, Luders, & Gaser, 2010;
MRIcro, 2020; Penny, Friston, Ashburner, Kiebel, & Nichols, 2011). ALVIN and VBM8 are an SPM8 plugin. MRICro is used for
preprocessing and visualization. The programs used are not sufficient for volume calculation alone. The output of one program can be

European Journal of Science and Technology
e-ISSN: 2148-2683 601
the input of another program. Therefore, in this study, a model regarding the order in which package programs and libraries can be
used is also proposed. For the analysis of AD disease, a total of twenty metrics were derived, with different combinations of six global
volume regions: GM, WM, CSF, LV, Total Intracranial Volume (TIV (GM+WM+CSF)), and Parenchyma (GM+WM). The measured
values were analyzed with independent samples t-test. It was evaluated whether the values were significant or not.
2. Material and Methods
2.1. Data Set
70 AD and 70 NC T1 weighted MR images were taken from the OASIS database. The OASIS database consists of MR images
ranging in age from 18 to 96. The images are composed of young and old AD MR images. MR data are labeled with Mini Mental
Status Test (MMSE) and Clinical Dementia Rating (CDR) neuropsychological tests. MMSE scoring ranges from 0-30, while CDR
ranges from 0-3. The OASIS database contains only labeled CDR0-2. Therefore, images labeled CDR0 as NC, images labeled
CDR0.5-2 were used as AD in this study. Since age 65 is a risk factor for AD, data above this age were taken from the database. Data
set used and demographic distrubution are given in Table 1. There are MR images already pre-processed and registered to the
Talairach axis in the database. However, these registered images do not represent actual volume values. Therefore, only the raw
images prefixed with "OAS1_xxxx_MRy_mpr _ni_ anon_sbj_111" were taken. The necessary pre-processing on the raw images have
been performed in the study. OASIS database imaging protocols are TE:4.0 msec, TR:9.7 msec, TI:20 msec, flip angle=10, 128
sagittal, 1.25 slice thickness without gap, pixel resolution 256x256 (1x1mm) and T1 weighted magnetization-prepared rapid gradient
echo (MPRAGE) multiple high- resolution images via 1.5 T vision scanner (Siemens, Erlangen, Germany)(Marcus et al., 2007).
Table 1: Data set used in this study and demografic distrubution
NC
AD
Age Gr.
Numb.
M\F
Mean Age
Mean MMSE
CDR
Age Gr.
Numb.
M\F
Mean Age
Mean MMSE
CDR
60s
9
3\6
68.00±1.11
28.77±1.64
0
60s
7
3\4
67.85± 1.21
23.42± 4.85
5\2\0
70s
32
8\24
73.37±2.44
29.15±0.91
0
70s
35
14\21
74.42±2.61
24.54±4.21
23\11\1
80s
22
6\16
83.40±3.27
28.77±1.26
0
80s
23
11\12
82.69± 2.61
24.47 ±4.06
16\6\1
90s
7
1\6
91.14±1.67
28.57±1.71
0
90s
5
2\3
92.00± 2.44
23.80±1.92
4\1\0
Total
70
18\52
77.61±7.48
28.92± 1.21
0
Total
70
30\40
77.74 ±6.66
24.35 ±4.05
48\20\2
2.2. Measurement of Global Brain Volumes
To make volume analysis in 3D raw MR images, some pre-processing techniques must be done. Firstly, images were converted
from the sagittal axis to axial axis with MRIcro program. Because the template images in VBM8 and SPM8 programs are defined on
the axial axis, this process must be done. With the "Display" option of SPM8 program, a reorientation process was performed to the
Anterior Commissure (AC) point. This point is considered to be the center point of the brain (x, y, z = 0) (Talaraich & Tournoux,
1988). This process is called AC/Posterior Commissure (PC) line correction. Otherwise, the MATLAB program gives an error in the
next steps. These data for the segmentation process were given to input of the VBM8 library. VBM8 library is an SPM8 plug-in that
combines the preprocessing and segmentation stages of structural MR images. Using this plug-in, bias correction, denoising,
normalization, and segmentation processes were performed automatically on the 3D structural MR images. The high-dimensional
Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL) method was chosen for normalization to the
Montreal Neurological Institute (MNI) space. The DARTEL method was proposed by John Ashburner (Ashburner, 2007). This
normalization method gives the best register among 14 different normalization algorithms (Klein et al., 2009). It also gives the best
register between MR images taken with different protocols from different centers (Goto et al., 2013). The Modulation method was
preferred for segmentation. Modulation operation is an option that maintains the volume in the native space before normalization
(Mechelli, Price, Friston, & Ashburner, 2005). As a result of the processes, segments of GM, WM ,and CSF regions were obtained in
Montreal Neurological Institue (MNI) space. Image volumes were calculated by using “get_totals.m” MATLAB script developed by
Ged Ridgway (Ridgway, 2020). LV volume and segmentation were obtained with the ALVIN program over the modulated CSF
images segmented with the SPM8 program (Kempton et al., 2011). The amount of TIV was calculated by summing up GM, WM and
CSF volumes, and the Parenchyma (PRM) tissue volume was calculated by summing up GM and WM volumes. Thus, the volumes of
six regions in total were calculated in milliliters (ml). In Figure 2, a flow diagram that performs pre-processings on raw 3D T1-
weighted MR image of the subject numbered 430 from OASIS database and finding segmentation and volume values of GM, WM,
CSF, LV, TIV, PRM regions are given. Three-dimensional models of four global brain regions are shown in Figure 3. In Figure 4, the
borders of the GM, WM, CSF, PRM, TIV regions on the MR image and the LV region's borders on the modulated segmented GM are
given.

Avrupa Bilim ve Teknoloji Dergisi
e-ISSN: 2148-2683 602
Figure 2. Flow diagram that performs pre-processings on raw 3D T1-weighted MR image of the subject numbered 430 from OASIS
database and finding segmentation and volume values of GM, WM, CSF, LV, TIV, PRM regions (Öziç, 2018)

European Journal of Science and Technology
e-ISSN: 2148-2683 603
Figure 3. Three-dimensional models of four global brain regions (GM, WM, CSF, LV)
(a) (b) (c) (d)
(e) (f) (g)
Figure 4. (a) Converted axial plane OASIS 430 raw MR image (b) GM borders on raw MR image (c) WM borders on raw MR image
(d) CSF borders on raw MR image (e) Parenchyma borders on raw MR image (f) TIV borders on raw MR image (g) LV borders on
modulated segmented GM
3. Research Results and Discussion
GM, WM, CSF, PRM, TIV, LV volumes were obtained by using the flow diagram in Figure 2 in each T1-weighted 3D MR image.
With the volumes of the six regions obtained, twenty-six metrics were derived from different ratios and combinations. Mean and
standard deviations of all metrics were calculated for AD and NC groups. Whether the values are significant between AD and NC was
calculated by independent samples t-test. Mean, standard deviation and p significance value of the metrics are given in Table 2. The
metrics were sorted in the table from the most significant to the most insignificant with respect to p value.

Avrupa Bilim ve Teknoloji Dergisi
e-ISSN: 2148-2683 604
Table 2: Mean, standard deviation and p significance values of the metrics(Öziç, 2018)
AD
NC
Stats
Mean
Std
Mean
Std
p
CSF/GM
0.598
0.085
0.53
0.072
1,00E-06
(WM+CSF)/TIV
0.586
0.022
0.57
0.016
2,00E-06
GM/TIV
0.414
0.022
0.43
0.016
2,00E-06
(LV+CSF)/TIV
0.28
0.034
0.253
0.033
7,00E-06
CSF/TIV
0.246
0.025
0.227
0.024
9,00E-06
PRM/TIV
0.754
0.025
0.773
0.024
9,00E-06
CSF/ PRM
0.328
0.044
0.295
0.041
1,00E-05
CSF
337
47.97
305.3
40.49
4,00E-05
LV/GM
0.082
0.032
0.063
0.027
0.0002
WM/CSF
1.Nis
0.19
1.538
0.231
0.0002
LV/ PRM
0.045
0.017
0.035
0.015
0.0004
LV/TIV
0.034
0.012
0.027
0.011
0.0006
LV
46.44
19.77
36.01
15.58
0.0007
LV/WM
0.099
0.038
0.078
0.034
0.0009
(GM+LV)/TIV
0.447
0.019
0.456
0.013
0.0018
LV/CSF
0.135
0.044
0.115
0.041
0.0068
WM/GM
0.824
0.064
0.801
0.051
0.0149
GM/ PRM
0.549
0.019
0.556
0.016
0.018
WM/ PRM
0.451
0.019
0.444
0.016
0.018
GM
565.9
55.93
578
47.35
0.1706
(WM+LV)/TIV
0.374
0.018
0.37
0.017
0.2347
(GM+CSF)/TIV
0.66
0.017
0.656
0.018
0.2548
WM/TIV
0.34
0.017
0.344
0.018
0.2548
TIV
1369
132.9
1346
116
0.2861
PRM
1032
104.7
1041
96.57
0.5934
WM
466.2
55.73
463.2
53.19
0.7488
The most significant value in AD and NC volume comparison is found in CSF/GM. GM/TIV and (WM+CSF)/TIV demonstrate a
high degree of statistical significance after CSF/GM. GM, (WM+LV)/TIV, (GM+CSF)/TIV, WM/TIV, WM, TIV, PRM values are
statistically insignificant. Since AD begins with forgetfulness, initial volumetric differences occur in memory-related brain regions on
GM. In the literature, there are studies showing that volumetric losses and differences occur in GM and WM using voxel-based
morphometry (G. Frisoni et al., 2002; Guo et al., 2010). But in this study, GM, WM, Parenchyma volume values alone gave
meaningless results. Even if volumetric loss occurs in these regions, the different head size of each person gives meaningless results in
volume comparisons. To overcome this problem, it is usually used by normalizing the measurements to a value. Normalization is
performed by dividing the measured value by TIV or GM. Another approach is the proportion of measured values to each other (Ge et
al., 2002; Orellana et al., 2016; Youn & Hsiung, 2015). The aim of these approaches, even if the values measured in healthy people
differ among themselves, their proportions to each other or global volumes will be similar. Therefore, patient-normal comparisons
made using volume values will yield more objective results. Since there is the volumetric loss in GM and WM regions, CSF is filled
to these regions. Since CSF increases too much, it shows a significant difference both alone and in TIV normalization. Significant
differences were observed in LV volume alone and in TIV normalization as the LV region expands and the volume increase is very
high (Bigler, 2015; Rababa'h, 2014). Since the skull does not shrink and the total brain volume does not change, it is an expected
result that no significant difference TIV volume is observed. Since the PRM region is the sum of WM and GM, the change in this
region significantly affected the statistical results in the combinations using PRM. As a result of the experiments, the normalization
process to the TIV and Parenchyma regions mostly gives meaningful results. At the same time, the proportion of areas where
volumetric loss is expected gives each other meaningful results. CSF/GM, (WM+CSF)/TIV, GM/TIV, (LV+CSF)/TIV, PRM/TIV,
CSF/PRM, LV/ GM, WM/CSF, LV/PRM, LV/GM, WM/CSF, LV/PRM, LV/TIV, LV/WM, (GM+LV)/TIV, LV/CSF, WM/GM,
GM/PRM, WM/PRM rates and normalization processes are the remarkable results of the study.

European Journal of Science and Technology
e-ISSN: 2148-2683 605
4. Conclusion
Volumetric differences measured for AD in 3D T1-weighted MR images have a biomarker feature for the disease. In this study,
global volumetric calculations were made with a model developed in AD and NC MR images, the ratios of regions to each other and
the results of normalization processes were evaluated statistically. In the literature, although values such as GM / WM, GM / TIV are
generally used for comparison, in this study has been determined that the ratios of other regions in the brain may give statistically
significant results. Therefore, different rates determined in the study can be used as a biomarker for the disease. The different head
sizes of each person can give misleading results for the analysis of regions alone in volumetric analysis studies. In this study,
misleading points are statistically revealed and discussed after volumetric measurements. CSF/GM ratio has been proven in the study
that it is an important and powerful biomarker for the disease. Manual interpretation of 3D MR images is a laborious, user-prone,
time-consuming process. Computer-aided numerical methods can give more accurate and faster measurement results practically. With
the measurement method used in this study, global brain regions can be measured quickly. In the future, it is an expected development
in the field of medicine that such numerical methods are analyzed through the MR device or server and directed to the doctors as a
pre-diagnosis in the form of a report.
Kaynakça
Ashburner, J. (2007). A fast diffeomorphic image registration algorithm. Neuroimage, 38(1), 95-113.
Ashburner, J., Barnes, G., Chen, C., Daunizeau, J., Flandin, G., Friston, K., . . . Litvak, V. (2008). SPM8 manual. Functional Imaging
Laboratory, Institute of Neurology, 41.
Association, A. s. (2019). 2019 Alzheimer's disease facts and figures. Alzheimer's & Dementia, 15(3), 321-387.
Bigler, E. D. (2015). Structural image analysis of the brain in neuropsychology using magnetic resonance imaging (MRI) techniques.
Neuropsychology Review, 25(3), 224-249.
Derneği, T. A. (2020). Türkiye'de 600bin aile Alzheimer Hastalığı ile Mücadele Ediyor. Retrieved from
http://www.alzheimerdernegi.org.tr/haber/turkiyede-600-bin-aile-alzheimer-hastaligi-ile-mucadele-ediyor/
Dubois, B., Hampel, H., Feldman, H. H., Scheltens, P., Aisen, P., Andrieu, S., . . . Blennow, K. (2016). Preclinical Alzheimer's disease:
definition, natural history, and diagnostic criteria. Alzheimer's & Dementia, 12(3), 292-323.
Fischl, B. (2012). FreeSurfer. Neuroimage, 62(2), 774-781.
Frisoni, G., Testa, C., Zorzan, A., Sabattoli, F., Beltramello, A., Soininen, H., & Laakso, M. (2002). Detection of grey matter loss in
mild Alzheimer's disease with voxel based morphometry. Journal of Neurology, Neurosurgery & Psychiatry, 73(6), 657-664.
Frisoni, G. B., Fox, N. C., Jack, C. R., Scheltens, P., & Thompson, P. M. (2010). The clinical use of structural MRI in Alzheimer
disease. Nature Reviews Neurology, 6(2), 67-77.
Ge, Y., Grossman, R. I., Babb, J. S., Rabin, M. L., Mannon, L. J., & Kolson, D. L. (2002). Age-related total gray matter and white
matter changes in normal adult brain. Part I: volumetric MR imaging analysis. American Journal of Neuroradiology, 23(8), 1327-
1333.
Goto, M., Abe, O., Aoki, S., Hayashi, N., Miyati, T., Takao, H., . . . Mori, H. (2013). Diffeomorphic Anatomical Registration Through
Exponentiated Lie Algebra provides reduced effect of scanner for cortex volumetry with atlas-based method in healthy subjects.
Neuroradiology, 55(7), 869-875.
Guo, X., Wang, Z., Li, K., Li, Z., Qi, Z., Jin, Z., . . . Chen, K. (2010). Voxel-based assessment of gray and white matter volumes in
Alzheimer's disease. Neuroscience letters, 468(2), 146-150.
Holland, D., Brewer, J. B., Hagler, D. J., Fennema-Notestine, C., Dale, A. M., Weiner, M., . . . Jagust, W. (2009). Subregional
neuroanatomical change as a biomarker for Alzheimer's disease. Proceedings of the National Academy of Sciences, 106(49),
20954-20959.
Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., & Smith, S. M. (2012). Fsl. Neuroimage, 62(2), 782-790.
Keller, S. S., & Roberts, N. (2009). Measurement of brain volume using MRI: software, techniques, choices and prerequisites. J
Anthropol Sci, 87, 127-151.
Kempton, M. J., Underwood, T. S., Brunton, S., Stylios, F., Schmechtig, A., Ettinger, U., . . . Frangou, S. (2011). A comprehensive
testing protocol for MRI neuroanatomical segmentation techniques: evaluation of a novel lateral ventricle segmentation method.
Neuroimage, 58(4), 1051-1059.
Klein, A., Andersson, J., Ardekani, B. A., Ashburner, J., Avants, B., Chiang, M.-C., . . . Hellier, P. (2009). Evaluation of 14 nonlinear
deformation algorithms applied to human brain MRI registration. Neuroimage, 46(3), 786-802.
Kurth, F., Luders, E., & Gaser, C. (2010). VBM8 toolbox manual. Jena: University of Jena.
Marcus, D. S., Wang, T. H., Parker, J., Csernansky, J. G., Morris, J. C., & Buckner, R. L. (2007). Open Access Series of Imaging
Studies (OASIS): cross-sectional MRI data in young, middle aged, nondemented, and demented older adults. Journal of cognitive
neuroscience, 19(9), 1498-1507.
Mechelli, A., Price, C. J., Friston, K. J., & Ashburner, J. (2005). Voxel-based morphometry of the human brain: methods and
applications. Current medical imaging reviews, 1(2), 105-113.
MRIcro. (2020). Retrieved from https://people.cas.sc.edu/rorden/mricro/mricro.html
Orellana, C., Ferreira, D., Muehlboeck, J.-S., Mecocci, P., Vellas, B., Tsolaki, M., . . . Simmons, A. (2016). Measuring global brain
atrophy with the brain volume/cerebrospinal fluid index: normative values, cut-offs and clinical associations. Neurodegenerative
Diseases, 16(1-2), 77-86.
Öziç, M., Ü. (2018). 3B Alzheimer MR Görütnülerinin Sınıflandırılmasında Yeni Yaklaşımlar. (Doktora Tezi), Selçuk Üniversitesi, Fen
Bilimleri Enstitüsü, Konya, Türkiye

Avrupa Bilim ve Teknoloji Dergisi
e-ISSN: 2148-2683 606
Penny, W. D., Friston, K. J., Ashburner, J. T., Kiebel, S. J., & Nichols, T. E. (2011). Statistical parametric mapping: the analysis of
functional brain images: Elsevier.
Petrella, J. R., Coleman, R. E., & Doraiswamy, P. M. (2003). Neuroimaging and early diagnosis of Alzheimer disease: a look to the
future 1. Radiology, 226(2), 315-336.
Petropoulos, H., Sibbitt Jr, W. L., & Brooks, W. M. (1999). Automated T2 quantitation in neuropsychiatric lupus erythematosus: a
marker of active disease. Journal of Magnetic Resonance Imaging, 9(1), 39-43.
Rababa'h, Q. (2014). Intracranial volume Segmentation. (Master Thesis in Medicine), Örebro University,
Ridgway, G. (2020). Miscellaneous useful MATLAB scripts for SPM/VBM. Retrieved from
http://www0.cs.ucl.ac.uk/staff/g.ridgway/vbm/get_totals.m
Salat, D. H., Greve, D. N., Pacheco, J. L., Quinn, B. T., Helmer, K. G., Buckner, R. L., & Fischl, B. (2009). Regional white matter
volume differences in nondemented aging and Alzheimer's disease. Neuroimage, 44(4), 1247-1258.
Schuff, N., Woerner, N., Boreta, L., Kornfield, T., Shaw, L., Trojanowski, J., . . . Initiative, D. N. (2009). MRI of hippocampal volume
loss in early Alzheimer's disease in relation to ApoE genotype and biomarkers. Brain, 132(4), 1067-1077.
Selekler, K. (2010). Alois Alzheimer ve Alzheimer Hastalığı. Türk Geriatri Dergisi, 13(3), 9-14.
Talaraich, J., & Tournoux, P. (1988). Co-planar stereotaxic atlas of the human brain. George Thieme, Stuttgard.
Vemuri, P., & Jack, C. R. (2010). Role of structural MRI in Alzheimer's disease. Alzheimer's research & therapy, 2(4), 23.
Villarreal, G., Hamilton, D. A., Petropoulos, H., Driscoll, I., Rowland, L. M., Griego, J. A., . . . Brooks, W. M. (2002). Reduced
hippocampal volume and total white matter volume in posttraumatic stress disorder. Biological psychiatry, 52(2), 119-125.
Youn, Y. C., & Hsiung, G.-Y. R. (2015). A Voxel Based Morphometric Analysis of Longitudinal Cortical Gray Matter Changes in
Progranulin Mutation Carriers At-Risk for Frontotemporal Dementia: Preliminary Study. Dementia and Neurocognitive
Disorders, 14(4), 163-167.