Desktop version

Home arrow Health

Brain Image Database

To assess normal variations of the human brain across populations of different ages and genders or to detect pathological differences between normal and diseased brains, large amounts of imaging data are necessary. Recently, projects to construct a large-scale database for human brain images have been vigorously pursued by major neuroscience centers in Europe and the USA.

There are many groups involved in neuroimaging database development. Among them, the International Consortium for Brain Mapping (ICBM) [127] is well known and is one of several powerful, multicenter groups. This consortium is governed by Professor John Mazziotta (UCLA) and is composed of four core research sites: UCLA, MNI, the University of Texas at San Antonio, and the Institute of Medicine, Juelich/Heinrich Heine University in Germany. In addition, data acquisition sites in Asia (Sendai, Japan, this author’s group) [128] and Europe (France, Finland, and the Netherlands) contribute to this international consortium.

We founded the Aoba Brain Imaging Center (ABIC) (with this author as the project leader) in1998. At the end of this project, ABIC had collected brain MRIs of 1600 healthy Japanese. We registered the brain images together with information on the subjects’ age, sex, height, weight, blood pressure, medical history, social history, and other factors to construct a database [244]. The number of cases at the time of writing is 2743 (Table 3.3).

Table 3.3 Number of subjects in healthy Japanese brain MRI database

Number of subjects

Sub-database

Male

Female

Total

Aoba-1

805

786

1591

Aoba-2

184

258

442

Tsurugaya-1

92

104

196

Tsurugaya-2

118

105

223

Children

145

146

291

Total

1344

1399

2743

All subjects were living in or around Sendai City, Japan. Age of the subjects ranged from 18 to 80 in Aoba-1 and -2, 70 or older in Tsrugaya-1 and -2, and from 6 to 18 in the Children sub-database

Application of VBM to Medical Imaging: Gray Matter Volume Loss in Patients with Sub-threshold Depression VBM has been used in studies on normal brain aging and the characterization of brain pathology. Good [108] first reported age- related volume changes in the human brain using VBM. Taki also analyzed age-related brain volume changes in a large number of healthy Japanese subjects in both cross-sectional [280, 282] and longitudinal [283] designs. Baron et al. [28], using VBM, demonstrated that patients with Alzheimer disease showed gray matter loss in medial temporal structures, the posterior cingulate gyrus, and the temporoparietal association cortex. Whitwell reported gray matter loss in patients with frontotemporal lobar dementia [311] and Parkinson’s disease [310]. In the following section, some results from our own studies will be introduced.

Depression is one of the most common psychiatric disorders in the elderly. There are subjects who have significant depressive symptoms, but who do not meet the criteria for major depression. However, the symptoms in these patients are associated with deterioration of physical function, worsening of physical diseases, a higher risk of mortality, and a higher risk of suicide, and thus, this state is considered to be depression syndrome and is clinically important in the elderly. It is termed “sub-threshold depression” (sD).

We assessed for differences in regional gray matter volume between communitydwelling elderly subjects with sD and age-matched nondepressed normal subject using VBM. We defined subjects with sD as those who scored > 15 on the Geriatric Depression Scale and > 22 on the Mini-Mental State Examination and did not fulfill the criteria for major depressive disorder (MDD) in the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV). We collected brain MRI data of 34 subjects with sD and 109 age-matched normal subjects and assessed differences in regional gray matter volume between these two groups by applying VBM.

Only male subjects with sD had significantly smaller volumes of the medial parts of both frontal lobes and the right precentral gyrus when compared with normal male subjects (Fig. 3.19) [281]. There were no significant structural differences between female subjects with sD and normal female subjects. Our study revealed

The areas in which gray matter volume decreases in sub-threshold depression Medial frontal lobe and right prefrontal gyrus

Fig. 3.19 The areas in which gray matter volume decreases in sub-threshold depression Medial frontal lobe and right prefrontal gyrus (Taki Y et al., J Affective Disorders 88:313-320, 2005. Courtesy of Elsevier. Figure 1 of Ref. [281])

that community-dwelling elderly male subjects with sD showed bilateral prefrontal gray matter volume reduction, which has been reported in elderly patients with MDD. Unlike findings in MDD, there was no significant volume reduction in the hippocampus. Our study implies the mechanism underlying the pathogenesis of brain volume loss and its relationship with sD in males.

Application of DBM in Computation of Standard Brain Models for Each Age and Sex Group DBM has been applied to assess brain shape changes in alcoholism [53], brain atrophy [102], and Alzheimer disease [88]. We sought to generate a Japanese standard brain model for each gender and age group (20-70 years old) using DBM in a large number of brain MRIs of healthy Japanese [244]. Subjects were obtained from the Aoba-1 database (Table 3.3) and were divided into age- sex groups for each decade (1920s, 1930s, 1940s, 1950s, 1960s, and 1970s). All were healthy, nondemented, and had no history of brain diseases. Abnormal brain MRIs, such as those showing cerebrovascular diseases, brain tumors, or massive white matter ischemic changes, were excluded.

Based on our previous study [244], we selected a brain MRI scan that showed the least sum energy in the process of linear transformation into other brains in each age-sex group. The brains in each age-sex group were used as reference brain (R) in this study. Each brain image T in each age-sex group was coregistered with image R of its group by a six-parameter affine transformation, and a coregistered image T0 was generated. T0 had the same shape and size as T. The deformation field, which transformed image R into image T0, was calculated as follows: (1) image R was roughly deformed to image T0 by a 12-parameter affine transformation, and transformed image R0 was generated. This affine transformation was converted to a deformation field A. (2) Image R0 was deformed to image T0 by an elastic transformation algorithm [250], and deformation field B was generated. (3) The composition of deformation fields A and B was calculated (C = BA) [245]. C contained both the 12-parameter affine transformation and a nonlinear elastic transformation. C was obtained for each brain image dataset of the age-sex group.

All the deformation field Cs in the age-sex group were averaged by calculating the mean deformation vector in each voxel and generated mean deformation field M. By applying M to the reference image R, a brain image representing the mean shape of the age-sex group was generated.

Deformation fields that transformed a reference brain from the 40s age group into the brains of other age groups were calculated. In this study, a brain of a 41- year-old male was used as R for aging simulation in male brains. The steps used in obtaining the transformation fields were the same as those used to calculate an averaged brain for each age group. All deformation fields were averaged; this was termed a “deformation field of aging.” Then, we computed a different age group brain by applying the deformation field of aging to a young reference brain (“age- simulated brain”).

Figure 3.20 shows computed averaged brains for each age group (left) and an age-simulated female brain (right) [94]. By visual inspection, ventricle and sulcus enlargement, specifically of the Sylvian fissure, were observed with increasing age. Male results were similar to those for females. Figure 3.21 (upper row) shows deformation fields that transformed a reference brain into other age group brains. Although the deformation field has three components for x, y, and z, only the

Averaged brains for each age group and age-simulated brains for each age group for males Left

Fig. 3.20 Averaged brains for each age group and age-simulated brains for each age group for males Left: averaged brains for each age. Right: age-simulated brains for each age. The images were shown on the plane parallel to that including the AC-PC line (Fukuda H et al., In Yamaguchi T ed.: Nano-biomedical Engineering 2012:170-190, Imperial College Press) (Figure 3 of Ref. [94])

S. Hanaoka et al.

186

Deformation matrices for aging simulation

Fig. 3.21 Deformation matrices for aging simulation. Deformation fields for aging simulation “40s!20s” indicate a transformation of a brain of a 41-year-old into a 20s averaged brain. Brain images at the level of C10 mm parallel to the AC-PC line are shown (Fukuda H et al., In Yamaguchi T ed.: Nano-biomedical Engineering 2012: 179-190, Imperial College Press) (Figure 4 of Ref. [94])

horizontal plane values are shown by arrows. Figure 3.21 (middlerow) demonstrates age-simulated brains at the + 10 mm intercommissural (AC-PC) line level for each age in males. A deformation field of aging transformed a young brain into an older brain with large ventricles and sulci. However, simulated aged brains were similar to averaged brains, although the transformation did not fully match the averaged brains (lower row).

Acknowledgments

Our studies were partly supported by a grant-in-aid from the Ministry of Education, Culture, Sports, Science and Technology (No 23,240,056, No 24,103,701).

 
Source
< Prev   CONTENTS   Source   Next >

Related topics