The Stereology and 3D Volume Analyses in Nervous Tissue - Animal Models of Acute Neurological Injuries: Injury and Mechanistic Assessments - page 44

a simple and popular approach for 3D analysis of neuroanatomical
images and is based on measuring the volume either of a structure
of interest or of a number of partitions of the brain structure.
For example, an MRI study using this method quantifi ed the
hippocampal volume from MRI data and identifi ed bilateral volu-
metric reduction of the hippocampus in schizophrenia ( 12 ).
The following is a simple method for estimating the volume
of an ROI: (1) count the number of the voxels in the ROI and
(2) multiply it by the voxel size. Given two or more groups of
subjects, statistical analysis can then be performed on their volume
measures by using standard packages such as SAS ( http://www.sas.
com/ ) , SPSS ( http://www.spss.com/ ), R ( http://www.r-project.
org/ ) , or Excel ( http://www.microsoft.com/ ) to quantify the
group difference.
The main advantage of volumetric analysis is its simplicity;
however, many structural differences may be overlooked. Newer
approaches, such as VBM (described in Section 7.2 ) and SBM
(see Section 7.3 ), have the potential to provide information beyond
simple volume measurements, and may characterize abnormalities
(e.g., small and regional changes) in the absence of volume
differences.

7.2. Registration
and VBM

Registration is another critical medical image computing technique,
which transforms different image data into the same coordinate
system in order to facilitate comparison across different image
subjects. VBM is a typical application where the registration
technique plays an important role. It is widely used in analyzing
neuroimaging data and allows specifi c tissue classes (e.g., gray
matter, white matter, or cerebrospinal fl uid) to be analyzed in an
automated and objective manner ( 17 ).
VBM involves voxel-by-voxel statistical analysis of the local
concentration of gray matter or white matter and can be applied to
compare two or more groups of subjects. For example, Fig. 5
shows the results of a sample VBM study using SPM (Table 3 ),
where signifi cant regions of gray matter loss were identifi ed in patients
with Alzheimer's disease compared with healthy controls ( 13 ).
The VBM protocol for examining difference across subjects by
analyzing MRI scans is standard. The following shows a typical
VBM processing pipeline using the SPM software package (Table 3 ):

(1) Launch SPM5 in Matlab (see its graphical user interface in
Fig. 9 ).
(2) Run “Coregister” option to roughly align all the individual
scans to a standard template image.
(3) Run “Segment” option to segment scans into gray matter,
white matter, and cerebrospinal fl uid compartments using
standard SPM5 templates.