JP4335736B2 - Brain image data processing system, method, program, and recording medium - Google Patents

Description

  The present invention relates to a brain image data processing system or the like that causes a computer to execute processing of brain image data using an artificial brain specimen in order to diagnose a brain function of a subject, and in particular, brain images acquired by different imaging devices in different facilities The present invention relates to a brain image data processing system or the like that causes a computer to execute data processing.

  In recent years, diagnostic imaging has advanced remarkably with the advance of medicine. An image diagnostic machine, such as an X-ray tomograph (CT), a magnetic resonance imaging machine (Magnetic), which captures the state of the body as an image and makes an accurate diagnosis without imposing a heavy burden on the patient. Resonance Imaging (MRI), ultrasonic diagnostic equipment, radiological diagnostic equipment, etc. are indispensable in the current medical field. Image diagnosis using the above-described image diagnostic machine is widely used as providing functional information such as early diagnosis of a disease of a patient, selection of a treatment method, prediction and determination of a treatment effect, and the like.

  In the clinical field of nuclear medicine, a single photon emission computed tomography (Radio Isotope: RI) is introduced into a patient's body and γ-rays emitted from the radioisotope (RI) are used. SPECT) and positron emission tomography (PET) are used by using the respective apparatuses. In the medical image processing apparatus as described above, various computer programs are prepared so that various image processing such as image reconstruction and image analysis can be performed from collected data.

  Recently, after converting the brain image data of each patient in the brain region on the stereotaxic standard brain chart score, by comparing it with the brain image database of healthy subjects, Attempts have been made to extract more accurately. At present, it is about to be applied in clinical practice as a standard brain coordinate system statistical analysis method. Specifically, by matching the brain images of each patient to the shape of a standard brain that “can identify anatomical position by coordinates”, it is possible to handle multiple brain images of different sizes and shapes in the same way Thus, comparisons between patients can be performed on the same pixel or three-dimensionally between voxels. In addition, by comparing the brain image data of each patient with a database of brain image data of healthy subjects (hereinafter referred to as “normal database” of brain images), an objective statistical analysis method can be used for abnormal sites. Can be extracted. As such a method, an SPM method (Statistical Parametric Mapping) and a 3D-SSP method (Three-Dimensional Stereotactic Surface Projections) are known. These methods were developed for the purpose of detecting an active region on a PET image, but have since been applied to a local cerebral blood flow analysis test in a SPECT image.

  Since the standard brain coordinate system statistical analysis method analyzes all pixels or voxels of an image, there is an advantage that the entire brain can be focused and there is no difference in the results by the analyst. On the other hand, when application in clinical diagnosis is taken into consideration, it becomes an n-to-1 statistical analysis using a plurality of healthy person data and one patient data. That is, in this case, it is necessary and most important to construct a normal database of standard brain images obtained by collecting a plurality of healthy person data. However, it is not only ethically difficult to collect brain function image data of healthy individuals with unified diagnostic imaging equipment, collection conditions, processing methods, etc. for each age and gender, but it also has an economic burden. Therefore, there is a problem that it is practically impossible in ordinary clinical facilities or general clinical hospitals.

  In order to solve the above-described problems, an image statistical analysis method (easy Z-score Imaging System: eZIS) of cerebral blood flow SPECT described in Non-Patent Document 1 and Non-Patent Document 2 has been developed. As described in the above-mentioned non-patent literature, in eZIS, (1) cerebral blood flow in which a brain phantom, which is an artificial brain specimen in the shape of a brain, is usually performed with imaging equipment in each facility Images are taken under the SPECT acquisition conditions. (2) In the hospital where the normal database of brain images is created, the same brain phantom is imaged under the same imaging equipment and acquisition conditions as when the normal database was created. The difference between the brain phantom image and the brain phantom image captured in the hospital that created the normal database is calculated for each pixel or voxel, and a correction matrix is created. (4) Using the correction matrix, A method of correcting a difference in brain image data due to an imaging device and collection conditions is used.

Usefulness of a New Image Statistical Analysis Method for Cerebral Blood Flow SPECT (easyZ-score Imaging System: eZIS), Hiroshi Matsuda, Monthly Inner Vision, November 2002 Sharing the image database of cerebral blood flow SPECT and the usefulness of 3D cerebral blood flow statistical analysis (eZIS), Hiroshi Matsuda, Nao Mizumura, revised in July 2002, published by Daiichi Radioisotope Institute

  As described above, by using eZIS, it is possible to correct a difference in brain image data due to an imaging device and a collection condition in each facility. As a result, it became possible to use a normal database of brain images between facilities.

  However, since the brain phantom is extremely expensive, there is a problem that it is substantially difficult to prepare the brain phantom at each facility. Furthermore, even if a brain phantom was available, the operation was complicated and not everyone could operate it. In addition, because the brain phantom is a model, not all facilities can cope with it because it cannot correct radiation absorption lines and scattered radiation accurately.

  Therefore, an object of the present invention is to solve the above problems, and an expensive brain phantom is used even in a clinical facility where it is difficult to collect a large number of brain image data of healthy persons and patients. It is an object of the present invention to provide a brain image data processing system or the like that can create an artificial brain specimen without using a normal database.

  The brain image data processing system according to the present invention is a brain image data processing system that causes a computer to execute processing of brain image data acquired by different imaging devices in different facilities. First normalizing means for performing anatomical normalization using a stereotaxic brain coordinate system on brain image data acquired by a first imaging device under a first condition for a plurality of healthy persons; The first brain image data recording means for recording the brain image data subjected to anatomical normalization by the converting means in a normal brain image database of the first facility, and the normal brain image database of the first facility A first average value calculating means for obtaining an average value for each pixel or voxel of the recorded brain image data, and using the average value obtained by the first average value calculating means, the health of the first facility A standard deviation calculating means for obtaining a standard deviation for each pixel or voxel of the brain image data recorded in the brain image database, and the computer of the second facility is configured to provide the first to at least one healthy person in the facility. Normalizing means for performing anatomical normalization using a stereotaxic coordinate system on brain image data acquired by a second imaging device different from the first imaging device under a second condition different from the first condition; A second brain image data recording means for recording the brain image data subjected to anatomical normalization by the second normalization means in a healthy person brain image database of a second facility; When there are a plurality of healthy persons, an average value for each pixel or voxel of the brain image data recorded in the brain image database of the healthy person in the second facility is obtained, and one healthy person in the second normalization means is obtained. A second average value calculating means for averaging a value for each pixel or voxel of the brain image data of the one person recorded in the brain image database of healthy persons of the second facility, and the second average value Artificial brain image data having the average value obtained by the calculation means and the standard deviation obtained by the standard deviation calculation means, and brain image data obtained by the second imaging device under the second condition for the subject Statistical analysis processing means for performing statistical analysis processing for comparing brain image data subjected to anatomical normalization using a stereotaxic coordinate system, and statistical analysis processing performed by the statistical analysis processing means. A display processing means for displaying on the display unit is provided.

  Here, in the brain image data processing system according to the present invention, the standard deviation calculation means uses the brain image data for each pixel or voxel recorded in the brain image database for healthy persons in the first facility, The standard deviation for each pixel or voxel of the brain image data can be obtained by resampling method (preferably bootstrap method but jackknife method or the like).

  Here, in the brain image data processing system of the present invention, the statistical resampling method in the standard deviation calculating means can be a bootstrap method.

  The brain image data processing method according to the present invention is a brain image data processing method that causes a computer to execute processing of brain image data acquired by different imaging devices of different facilities, wherein the computer of the first facility A first normalization step of performing anatomical normalization using a stereotaxic coordinate system on brain image data acquired by a first imaging device under a first condition for a plurality of healthy persons; and the first normalization The first brain image data recording step for recording the brain image data subjected to the anatomical normalization in the step into the healthy person brain image database of the first facility, and the brain image data recorded in the healthy person brain image database of the first facility A first average value calculating step for obtaining an average value for each pixel or voxel of the brain image data obtained, and using the average value obtained in the first average value calculating step, A standard deviation calculating step for obtaining a standard deviation for each pixel or voxel of the brain image data recorded in the brain image database of the healthy person, and a computer of the second facility for at least one healthy person in the facility Second normalization for performing anatomical normalization using a stereotaxic coordinate system on brain image data acquired by a second imaging device different from the first imaging device under a second condition different from the first condition A second brain image data recording step for recording the brain image data subjected to anatomical normalization in the second normalization step in a normal brain image database of a second facility; and the second normalization When there are a plurality of healthy persons in the step, an average value for each pixel or voxel of the brain image data recorded in the healthy person brain image database of the second facility is obtained, and the second normalized scan When there is one healthy person in the top, the second average value having the average value of each pixel or voxel of the brain image data of the one person recorded in the healthy person brain image database of the second facility Artificial brain image data having a calculation step, an average value obtained in the second average value calculation step and a standard deviation obtained in the standard deviation calculation step, and a second image of the subject under the second condition Statistical analysis processing step for performing statistical analysis processing for comparing brain image data obtained by anatomical normalization using a stereotaxic brain coordinate system with respect to brain image data acquired by a device, and the statistical analysis processing step And a display processing step for displaying the statistical analysis processing performed in step 1 on a display unit.

  Here, in the brain image data processing method of the present invention, the standard deviation calculating step uses a brain image data for each pixel or voxel recorded in the brain image database of healthy persons in the first facility, The standard deviation for each pixel or voxel of the brain image data can be obtained by resampling method (preferably bootstrap method but jackknife method or the like).

  Here, in the brain image data processing method of the present invention, the statistical resampling method in the standard deviation calculating step can be a bootstrap method.

  The brain image data processing program of the present invention is a brain image data processing program for causing a computer to execute processing of brain image data acquired by different imaging devices in different facilities, and the computer is operated by at least one healthy person in the facility. In contrast, normalization means for performing anatomical normalization using a stereotaxic brain coordinate system on brain image data acquired by a predetermined imaging device under predetermined conditions, and anatomical normalization is performed by the normalization means Brain image data recording means for recording the brain image data in the healthy person brain image database of the facility, and when there are a plurality of healthy persons in the normalizing means, the brain image data recorded in the healthy person brain image database of the facility An average value for each pixel or voxel is obtained, and when there is one healthy person in the normalizing means, the brain image database of healthy persons in the facility Mean value calculating means for taking the average value of each pixel or voxel of the recorded brain image data of the one person, the average value obtained by the average value calculating means, and a computer of another facility different from the facility By using a stereotaxic coordinate system for brain image data acquired by another imaging device different from the predetermined imaging device under other conditions different from the predetermined condition for a plurality of healthy persons in the other facility The average for each pixel or voxel of brain image data that has been subjected to anatomical normalization and has been subjected to the anatomical normalization and recorded in the brain image database of the healthy subject of the other facility An artificial brain image data having a standard deviation obtained for each pixel or voxel of brain image data recorded in the healthy subject brain image database of the other facility using the average value. Statistical analysis of the brain image data obtained by performing anatomical normalization using a stereotaxic brain coordinate system on the brain image data acquired by the predetermined imaging device under the predetermined condition for the subject A statistical analysis processing means for performing processing, and a program for causing the statistical analysis processing performed by the statistical analysis processing means to function as display processing means for displaying on a display unit.

  Here, in the program according to the present invention, the standard deviation calculating means uses a statistical resampling method (using a brain image data for each pixel or voxel recorded in the brain image database of the healthy person in the first facility). The standard deviation for each pixel or voxel of the brain image data can be obtained by a bootstrap method (or a jackknife method or the like).

  Here, in the program of the present invention, the statistical resampling method in the standard deviation calculating means can be a bootstrap method.

  The recording medium of the present invention is a computer-readable recording medium on which any of the programs of the present invention is recorded.

  According to the brain image data processing system and the like of the present invention, first, each pixel or voxel of brain image data is obtained from a brain image database (normal brain image database) of a healthy person acquired in a facility that satisfies the first condition. A standard deviation for each pixel is obtained, and an average value for each pixel or voxel is obtained for brain image data of one healthy person acquired in another facility that satisfies the second condition. Next, the brain of the subject in the other facility is obtained by using an artificial brain sample in which the standard deviation of the healthy subject sample satisfying the first condition is fused with the data (average value) of any healthy person in the other facility. Performs statistical analysis of image data. As described above, the brain image data of the healthy person in the normal database of the brain image obtained under the first condition different from the acquisition condition (second condition) of the brain image data of the subject in the other facility is compared. The brain function of the subject can be diagnosed. For this reason, even in a clinical facility where it is difficult to collect a large amount of brain image data of healthy subjects and patients, an artificial brain specimen can be created without using an expensive brain phantom, and as a result, a normal database is used. It is possible to provide a brain image data processing system that can be used. Furthermore, it is possible to create a brain image database using the specialties peculiar to clinical facilities in the other facilities. By constructing a brain image database with few cases such as children and rare degenerative diseases that were difficult to create, there is an effect that it can be applied to the creation of brain coordinate samples by disease.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

The outline of the present invention is as follows.
(1) First, anatomical normalization using a stereotaxic coordinate system is performed on the brain image data of a healthy person acquired in a facility that satisfies the imaging device of the predetermined brain image data and the acquisition condition (first condition). , Record in the brain image database of normal subjects (normal database of brain images). Next, a standard deviation for each pixel or voxel of the brain image data recorded in the normal database of the brain image is obtained.
(2) Anatomical normality using a stereotaxic coordinate system for brain image data of one healthy person acquired in another facility that satisfies the imaging device and acquisition condition (second condition) of predetermined brain image data And an average value for each pixel or voxel is obtained for the brain image data.
(3) Artificial brain image data (artificial brain specimen) having the average value obtained in the above (2) and the standard deviation obtained in the above (1) is created.
(4) A comparison of the anatomical normalization of the brain image data of the subject in the other facility that satisfies the second condition and the artificial brain image data obtained in (3) above. Analytical processing is performed. It is a characteristic point of the present invention that the artificial brain specimen is used as a comparative control group. As described above, the brain image data of the healthy subject in the normal database of brain images obtained under the first condition different from the acquisition condition (second condition) of the subject's brain image data is used as a comparison target. The function can be diagnosed.

  The validity of using the artificial brain specimen having the above-described characteristics in the statistical analysis process has been verified by earnest research by the present inventors.

  The anatomical normalization and statistical analysis processing using the above-mentioned stereotaxic coordinate system is preferably performed using the SPM method or the 3D-SSP method. Anatomical normalization refers to performing position, size, and local deformation on the standard brain to eliminate anatomical differences in each individual's brain. In the case of SPECT or PET, in general, (1) the midline sagittal section of the brain in the SPECT or PET image is identified, and (2) four reference points in the same section (frontal pole, all lower end of corpus callosum, lower end of thalamus) And AC-PCline (line connecting the front and rear and rear commissures) as a reference line from (the posterior pole) and (3) moving to the coordinate system of the standard brain (Talairach atlas), (4) linear and nonlinear transformation This is done by going and deforming to the standard brain. As the pixel or voxel (pixel) data of the comparison control group in the statistical analysis process, it is preferable to use a group selected from the healthy group or the same disease group.

  Hereinafter, the image processing system of the present invention will be specifically described. FIG. 1 shows an image processing system of the present invention. In FIG. 1, reference numeral 10 denotes an entire image processing system including a facility N (first facility) 20 and a facility X (second facility) 40. In the facility N20, reference numeral 24 denotes an image processing apparatus (first facility computer) such as a personal computer (hereinafter referred to as “PC”) that executes the brain image data processing program of the present invention, and 26 denotes a brain image in the facility N20. An imaging device 28 for imaging / acquiring data is a normal database of brain images in which brain image data of a plurality of healthy persons are recorded. Reference numeral 30 is a block diagram showing the brain image data processing program of the present invention executed by the image processing device 24 in functional units, and 32 is acquired by the imaging device 26 under a first condition for a plurality of healthy persons in the facility N20. A first normalization unit (first normalization means) 34 that performs anatomical normalization using a stereotaxic coordinate system for brain image data, and 34 is a brain that has undergone anatomical normalization by the first normalization unit 32. A first brain image data recording unit (first brain image data recording means) for recording the image data in the normal database 28 of brain images of the facility N20 (the normal brain image database of the first facility), 36 is the brain of the facility N20 A first average value calculation unit (first average value calculation means) 38 for obtaining an average value for each pixel or voxel of brain image data recorded in the normal database 28 for images is obtained by the first average value calculation unit 38. By using the average value, a standard deviation calculation unit that calculates each pixel or standard deviation for each voxel of the brain image data recorded in the normal database 28 of the brain image (standard deviation calculation means).

  Subsequently, in the facility X40 shown in FIG. 1, reference numeral 44 denotes an image processing apparatus (computer of the second facility) that executes the brain image data processing program of the present invention, and 44 denotes input / output to / from the image processing apparatus 44. A terminal such as a PC to perform, 46 is an imaging device that captures / acquires brain image data in the facility X40, and 48 is a brain image database that records brain image data of at least one healthy person. Reference numeral 50 is a block diagram showing the brain image data processing program of the present invention executed by the image processing apparatus 44 in functional units, and 52 is acquired by the imaging device 46 under the second condition for one healthy person in the facility X40. The second normalization unit (second normalization means) 54 performs anatomical normalization on the brain image data using the stereotaxic coordinate system, and the second normalization unit 52 performs anatomical normalization. A second brain image data recording unit (second brain image data recording means) for recording the brain image data in the brain image database 48 of the facility X40 (the brain image database of the healthy person of the second facility), 56 is the brain image of the facility X40 A second average value calculating unit (second average value calculating means) for calculating an average value for each pixel or voxel of the brain image data recorded in the database 48, 58 is an average value obtained by the second average value calculating unit 56 And standard Anatomical normalization using a stereotaxic coordinate system for the artificial brain image data having the standard deviation obtained by the difference calculation unit 38 and the brain image data acquired by the imaging device 46 under the second condition for the subject. A statistical analysis processing unit (statistical analysis processing means) for performing a statistical analysis process for comparing the brain image data that has been performed, and a terminal 60 such as a PC for performing the statistical analysis process performed by the statistical analysis processing unit 58 44 is a display processing unit (display processing means) for displaying on the display unit 45.

  As indicated by the dotted line from the standard deviation calculation unit 38 in the facility N20 to the statistical analysis processing unit 58 in the facility X40, the standard deviation of the artificial brain image data used in the facility X40 is obtained by the standard deviation calculation unit 38 in the facility N20. Standard deviation is used. In the statistical analysis processing, for example, the statistical analysis result obtained by the following Expression 1 is performed on all pixels (in the case of three dimensions, voxels).

Statistical analysis result = (average artificial brain specimen pixel value−subject pixel value) /
Artificial brain specimen standard deviation (Formula 1)

  Here, as the artificial brain specimen standard deviation, the standard deviation obtained by the standard deviation calculation unit 38 in the facility N20 as described above is used.

  FIG. 2 is a flowchart showing the flow of the brain image data processing program or method of the present invention. In FIG. 2, the blocks from step S10 to S16 surrounded by a dotted line indicated by the symbol N indicate processing in the facility N20, and the blocks from steps S20 to S30 surrounded by the dotted line indicated by the symbol X are in the facility X40. Indicates processing.

  As shown in FIG. 2, the image processing apparatus 24 first uses an anatomical system using a stereotaxic brain coordinate system for brain image data acquired by the imaging device 26 under a first condition for a plurality of healthy persons in a facility N20. Normalization is performed (first normalization step, step S10). Next, the brain image data that has undergone anatomical normalization in the first normalization step (step S10) is recorded in the normal database 28 of the brain image of the facility N20 (first brain image data recording step, step S12). . Subsequently, an average value for each pixel or voxel of the brain image data recorded in the normal database 28 of the brain image of the facility N20 is obtained (first average value calculating step, step S14). Using the average value obtained in the first average value calculating step (step S14), a standard deviation for each pixel or voxel of the brain image data recorded in the normal database 28 of the brain image is obtained (standard deviation calculating step. Step S16).

  Subsequently, as illustrated in FIG. 2, the image processing device 44 is configured to detect at least one healthy person in the facility X40 with different characteristics from the imaging device 26 under a second condition different from the first condition. Anatomical normalization using the stereotaxic coordinate system is performed on the brain image data acquired by 46 (second normalization step, step S20). Next, the brain image data subjected to anatomical normalization in the second normalization step (step S20) is recorded in the brain image database 48 of the facility X40 (second brain image data recording step, step S22). Subsequently, an average value for each pixel or voxel of the brain image data recorded in the brain image database 48 of the facility X40 is obtained (second average value calculating step, step S24). Artificial brain image data having the average value obtained in the second average value calculation step (step S24) and the standard deviation obtained in the standard deviation calculation step (step S16), and the imaging device under the second condition for the subject Statistical analysis processing is performed to compare the brain image data acquired by 48 with brain image data that has been subjected to anatomical normalization using a stereotaxic coordinate system (statistical analysis processing step, step S28). The statistical analysis process performed in the statistical analysis process step (step S28) is displayed on the display unit 45 of the terminal 44 such as a PC (display process step, step S30).

The healthy person data used in the above-described image processing system 10, image processing program or method preferably satisfies the following categories (requirements). In addition to these requirements, categories such as gender, age, working environment or race may be used.
(1) Mini Mental State Examination (MMSE), Hasegawa simplified intelligence evaluation scale revision (HDS-R) is normal.
(2) Wechsler Memory Scale-Revised (WMS-R), Wechsler Adult Intelligence
Scale = Revised (WAIS-R) is normal.
(3) A high signal corresponding to the age is only seen in the white matter in the T2-weighted image by MRI.
(4) There are no risk factors for cerebrovascular disorders such as hypertension and diabetes.

The brain image data of a plurality of healthy persons at the facility N20 in the first normalization step (step S10) described above can be acquired, for example, as follows, and anatomical normalization can be performed. A first brain image data recording step (step S12), a first average value calculating step (step S14), and a standard deviation calculating step (step S16) are also shown.
(1) A drug is administered to a plurality of subjects who are considered to be healthy persons who have cleared the above category under the first condition, and imaging and reconstruction processes are performed by the imaging device 26 to create brain SPECT image data.
(2) A healthy person is determined by the brain SPECT image data created in (1) and other examinations.
(3) The brain SPECT image data of n healthy subjects determined in (2) is transferred from the imaging device 26 to the image processing device 24.
(4) The transferred n examples of brain SPECT image data are converted into an image data format called “Analyze format”.
(5) Each of the n examples of brain SPECT image data (brain SPECT image data having an individual-specific shape) created in (4) is converted into a standard brain coordinate system.
(6) Smoothing (smoothing) processing and count normalization for the brain image data created in (5) (eg, multiplying the brain image data by one scale factor so that the average count of the whole brain becomes 1) The brain image data (brain image specimen) of the healthy person is created on the brain image normal database 28 (first brain image data recording step, step S12). Smoothing processing refers to applying a normal distribution smoothing filter.
(7) The average value and standard deviation for each pixel (pixel or voxel) are obtained from the n examples of brain surface extracted image data created in (6), and the average brain image data and standard deviation brain image data are combined. Brain image data of healthy individuals in pairs is created (first average value calculating step (step S14), standard deviation calculating step (step S16)).

The brain image data of a plurality of healthy persons in the facility N20 described above can be acquired by the 3D-SSP method, for example, as follows. The first average value calculating step (step S14) and the standard deviation calculating step (step S16) are also shown.
(1) Under the first condition, a plurality of subjects who are considered to be healthy subjects are imaged and reconstructed by the imaging device 26 to create brain SPECT image data.
(2) A healthy person is determined by the brain SPECT image data created in (1) and other examinations.
(3) The brain SPECT image data of n healthy subjects determined in (2) is transferred from the imaging device 26 to the image processing device 24.
(4) The transferred n examples of brain SPECT image data are converted into an “3D-SSP format” image data format.
(5) Each of the n examples of brain SPECT image data (brain SPECT image data having an individual-specific shape) created in (4) is converted into a standard brain coordinate system.
(6) Brain table extraction processing is performed on the n examples of brain image data created in (5), and the above-described count normalization is performed.
(7) The average value and standard deviation for each pixel (pixel or voxel) are obtained from the n examples of brain surface extracted image data created in (6), and the average brain image data and standard deviation brain image data are combined. Brain image data of healthy individuals in pairs is created (first average value calculating step (step S14), standard deviation calculating step (step S16)).

Brain image data of a healthy person as an example of the facility X40 in the second normalization step (step S20) described above can be acquired as follows, for example, and anatomical normalization can be performed. This also includes the second average value calculation step (step S24).
(1) A drug is administered to one subject who is considered to be a healthy person who has cleared the category under the second condition, and imaging and reconstruction processing are performed by the imaging device 46 to create brain SPECT image data.
(2) A healthy person is determined by the brain SPECT image data created in (1) and other examinations.
(3) The brain SPECT image data of one example of a healthy person determined in (2) is transferred from the imaging device 46 to the image processing device 44.
(4) The transferred brain SPECT image data is converted into an image data format “Analyze format”.
(5) Each example of the brain SPECT image data (brain SPECT image data having a unique shape) created in (4) is converted into a standard brain coordinate system.
(6) Smoothing (smoothing) processing and normalization of the count are performed on the brain image data created in (5), and brain image data (brain image specimen) of a healthy person is created on the brain image database 48 ( Second brain image data recording step (step S22).
(7) The average value for each pixel (pixel or voxel) is obtained from the brain surface extracted image data of the example created in (6), and average brain image data is created (second average value calculating step (step S14) )).

The brain image data of the subject at the facility X40 in the statistical analysis processing step (step S28) described above can be acquired, for example, as follows, and anatomical normalization can be performed.
(1) Under the second condition, a drug is administered to the subject, and imaging and reconstruction processing are performed by the imaging device 46 to create brain SPECT image data.
(2) Transfer the brain SPECT image data of the subject from the imaging device 46 to the image processing device 44.
(3) The transferred brain SPECT image data of the subject is converted into an image data format called “Analyze format”.
(4) The subject's brain SPECT image data (brain SPECT image data having an individual-specific shape) created in (3) is converted into a standard brain coordinate system.
(5) Smoothing (smoothing) processing and count normalization are performed on the brain image data created in (4) to create brain image data of the subject.

  Hereinafter, an experimental example in the brain image data processing system of the present invention will be described.

Experimental Example 1
(1) In the facility N20, data for preparing a healthy specimen that seems to be a healthy person who cleared the above-mentioned category was collected with the following collection and imaging conditions and reconstruction processing conditions (first condition) being the same.
Acquisition and imaging conditions:
・ Camera model, collimator, energy window, sampling angle, collection direction, collection time, collection matrix, collection magnification, pixel size.
Reconfiguration processing conditions:
・ Preprocessing filter, reconstruction filter, scattered radiation removal, absorption correction.

(2) On the other hand, in the facility X40 to be diagnosed, brain image data of a healthy person was created as in (1). However, the collection and imaging conditions and the reconstruction processing conditions were the conditions (second conditions) normally implemented by the facility X40.
(3) The standard deviation for each pixel or voxel is obtained from the healthy subject sample created at the facility N20.
(4) An average value for each pixel or voxel is obtained from the healthy subject sample created at the facility X40.
(5) An artificial brain specimen is created by fusing the standard deviation obtained in (3) and the average value obtained in (4) for each pixel or voxel. Here, the fusion means that the standard deviation obtained at the facility N20 is used as the standard deviation of the artificial brain specimen at the facility X40.

Experimental Example 2.
The inventors of the present application verified the validity of the brain image processing system of the present invention using the following brain image data captured by different imaging devices 26 and 46 at the facility N20 and the facility X40.
Facility N20: 20 brains using 99mTc-ECD (“Neurolite (registered trademark) Injectable Solution 1” Daiichi Radioisotope Research Institute (registered trademark)) as a drug for healthy subjects who have cleared the above categories image data.
Facility X40: 15 cases of brains using 99mTc-ECD (“Neurolite (registered trademark) Injectable Solution 1” Daiichi Radioisotope Research Institute (registered trademark)) as a drug for healthy subjects who have cleared the above categories image data. In addition, brain image data of one example using 99mTc-ECD (“Neurolite (registered trademark) injectable solution first” First Radioisotope Laboratory (registered trademark)) as a drug for the subject.

Verification items.
(1) A statistical analysis process is performed using specimens of healthy persons (15 examples of the brain image database 48) of the same facility X40 as the subjects of the facility X40 (see FIG. 3).
(2) A statistical analysis process is performed using a specimen of a healthy person at a facility N20 (20 examples of the brain image normal database 28) different from the subject at the facility X40 (see FIG. 4).
(3) The artificial brain in which the standard deviation of the healthy subject specimens (20 examples of the normal database 28 of brain images) of the different facility N20 is fused to the subject of the facility X40 and any healthy subject data A of the same facility X40 Statistical analysis is performed using sample A (see FIG. 5).
(4) Artificial brain obtained by fusing the standard deviations of healthy subject specimens (20 examples of the normal database 28 of brain images) from different subjects N20 to the subject X40 subject and any healthy subject data B in the same facility X40 Statistical analysis is performed using the sample B (see FIG. 6).

Analysis result.
The result is a RENDER display method and a slice image, and P-value = 0.05 (P <0.5 ^n, where n is the number of categories) is used as a probability density to show statistical anomaly values (FIG. 3 to FIG. 3). (See FIG. 6). As shown in FIGS. 3 to 6, the display processing unit 60 of the brain image data processing system or the like can identify the area of the portion where the probability density P is a predetermined value or less from the result of the statistical analysis processing from other portions. For example, it can be displayed in one color. FIG. 3 shows a result of “same conditions” obtained by subjecting a subject at the facility X40 to a statistical analysis process using a healthy subject sample at the same facility X40 as a control group, that is, a correct statistical analysis process result. Fig. 3 (A) shows RENDER display (brain surface projection display, 3D). From the upper left side of Fig. 3 (A), the lower surface (Inferior), right side (R-lateral), and rear side of the brain. (Posterior), the left inner surface (L-medial) is displayed, and from the left side of the lower part of FIG. 3 (A), the upper surface of the brain (Superior), left side (L-lateral), front (Anterior), right The inside surface (R-medial) display is shown. As shown in FIG. 3 (A), a blood flow reduction site (partly dark part in the brain) was seen in the bilateral parietal cortex. FIG. 3B shows a Transaxial image display (two-dimensional). The 7 × 5 display in FIG. 3B is from the upper left to the lower right, and the body axis from the top to the bottom of the brain. The display of the image analysis result sliced in parallel is shown. As shown in FIG. 3 (B), blood flow decreased sites (partially dark portions in the brain) were seen in the bilateral parietal cortex.

  FIG. 4 shows a result of “different conditions” obtained by subjecting a subject at facility X40 to statistical analysis processing using healthy subject samples at different facilities N20 as a control group, that is, analysis without applying the brain image data processing method of the present invention. The statistical analysis processing results are shown. 4A shows a RENDER display similar to FIG. 3A, and FIG. 4B shows a Transaxial image display similar to FIG. 3B. Each display portion shows the same display results as those shown in FIGS. 3A and 3B, and a description thereof will be omitted. As shown in FIGS. 4 (A) and (B), the blood flow lowering sites of the bilateral parietal cortex are almost the same. However, the location indicated by P1 on the lower surface (Inferior), the location indicated by P2 on the upper surface (Superior), and the location indicated by P3 on the front surface (Anterior) in FIG. As is clear from the above, false positive statistical abnormal values are shown in the left temporal lobe and occipital lobe.

  FIG. 5 shows the result of the “fusion condition” analyzed by an artificial brain specimen obtained by fusing the standard deviation of the healthy subject specimen at the facility N20 to the data (average value) of an arbitrary healthy person A at the facility X40, that is, the brain image of the present invention. Statistical analysis processing results analyzed by applying a data processing method and the like are shown. 5A shows a RENDER display similar to FIG. 3A and the like, and FIG. 5B shows a Transaxial image display similar to FIG. 3B and the like. Each display portion shows the same display results as those in FIGS. 3A and 3B, and the description thereof is omitted. As shown in FIGS. 5A and 5B, it is clear that the blood flow reduction site in the bilateral parietal cortex coincides with the blood flow reduction site in FIG. From the above results, it was shown that the brain image data processing method of the present invention is sufficiently applicable to clinical practice.

  FIG. 6 shows the result of the “fusion condition” analyzed by an artificial brain specimen obtained by fusing the standard deviation of the healthy subject specimen at facility N20 to the data (average value) of an arbitrary healthy person B at facility X40, that is, the brain image of the present invention. Statistical analysis processing results analyzed by applying a data processing method and the like are shown. 6A shows a RENDER display similar to FIG. 3A and the like, and FIG. 6B shows a Transaxial image display similar to FIG. 3B and the like. Each display portion shows the same display results as those in FIGS. 3A and 3B, and the description thereof is omitted. As shown in FIGS. 6 (A) and 6 (B), even when the data (average value) of healthy subjects are exchanged, the blood flow reduction site in the bilateral parietal cortex matches the blood flow reduction site in FIG. It is clear that From the above results, it was shown that the brain image data processing method of the present invention is sufficiently applicable to clinical practice.

  As described above, according to the first embodiment of the present invention, each of the brain image data is first obtained from the brain image database (normal database of brain images) of the healthy person acquired in the facility that satisfies the first condition. A standard deviation for each pixel or voxel is obtained, and an average value for each pixel or voxel is obtained for brain image data of one healthy person acquired in another facility that satisfies the second condition. Next, the brain of the subject in the other facility is obtained by using an artificial brain sample in which the standard deviation of the healthy subject sample satisfying the first condition is fused with the data (average value) of any healthy person in the other facility. Performs statistical analysis of image data. As described above, the brain image data of the healthy person in the normal database of the brain image obtained under the first condition different from the acquisition condition (second condition) of the brain image data of the subject in the other facility is compared. The brain function of the subject can be diagnosed. For this reason, even in a clinical facility where it is difficult to collect a large amount of brain image data of healthy subjects and patients, an artificial brain specimen can be created without using an expensive brain phantom, and as a result, a normal database is used. It is possible to provide a brain image data processing system or the like that can be used.

  In the brain image data processing method and the like according to the present invention, the standard deviation calculating step (step S16) includes each pixel recorded in the normal database 28 (normal brain image database) of the brain image of the facility N10 (first facility). Using the brain image data for each voxel, the standard deviation for each pixel or each voxel of the brain image data is obtained by a statistical resampling method (preferably a bootstrap method but may be a jackknife method or the like). be able to. Statistical resampling method is a method in nonparametric statistical inference. The jackknife method is one of the statistical resampling methods, in which n-1 subsamples are extracted from n samples, and the corresponding statistics are calculated to improve the accuracy of estimation. The bootstrap method is also one of the statistical resampling methods.

  The following method was used as a simulation method such as the brain image data processing method of the present invention. That is, 12 sample groups were selected, and the prime samples were randomly extracted one by one from the 200 cases, and 12 sample groups were determined by repeating 12 steps. Using the sample group of 12 cases, the characteristics of 12 cases were compared with the characteristics resampled by the bootstrap method. The bootstrap method was performed according to the following procedure. However, the bootstrap method in the simulation method such as the brain image data processing method of the present invention is not limited to the procedure shown below, but is a parametric bootstrap method, a nonparametric bootstrap method, a semiparametric bootstrap. The method, the bootstrap-t method, the bootstrap iteration method, etc. can also be used.

procedure.
(1) Randomly extract and record one example of data from the twelve sample groups (data) determined as described above, and restore the data (restoration extraction).
(2) The above operation (1) is repeated 12 times to form one group. This operation is performed 200 times to create 200 groups of n = 12.
(3) A median is obtained for each group.
(4) The average value of the median values calculated for each group is obtained, and the standard deviation is calculated.

  As described above, normalization of sample groups can be artificially complemented by using the applicable bootstrap method even when the distribution or structure of statistics is not explicitly obtained. It is possible to improve the mutual correctness rate by adjusting the balance between one type of error and two types of error when verifying an event.

  FIG. 7 is a block diagram showing an internal circuit 70 of the computer of the image processing apparatus 44 that executes the computer program of the present invention for realizing Example 1 or 3 (described later) of the present invention. Since the internal circuit of the computer of the image processing device 24 is also shown in the same manner, only the image processing device 44 side will be described and the description of the image processing device 24 side will be omitted. As shown in FIG. 7, a CPU 72, a ROM 74, a RAM 76, an image control unit 78, a controller 86, an input control unit 90, and an external interface (Interface: I / F) unit 92 described later are connected to a bus 82. In FIG. 7, the above-described computer program of the present invention is recorded on a recording medium (including a removable recording medium) such as a ROM 74, a disk 84a, or a CD-ROM 84n. The computer program is loaded into the RAM 76 from the ROM 74 via the bus 82 or from a recording medium such as the disk 84a or CD-ROM 84n via the controller 86 via the bus 82. The image control unit 78 sends brain image data to the VRAM 80, and the display unit 45 is a display device such as a display that displays a brain image based on the brain image data sent from the VRAM 80. The VRAM 80 is an image memory having a capacity corresponding to the data capacity of one screen of the display unit 45. The input operation unit 88 is an input device such as a mouse or a numeric keypad for inputting to the image processing device 44. The input control unit 90 is connected to the input operation unit 88 and performs input control and the like. The external I / F unit 92 has an interface function when connecting to an external communication network (not shown) such as the Internet. The standard deviation obtained by the standard deviation calculation unit 38 of the image processing device 24 at the facility N20 can be used by the statistical analysis processing unit 58 of the image processing device 44 at the facility X40 via a communication network such as the Internet. Alternatively, the standard deviation is recorded on the removable recording medium 84n or the like of the image processing apparatus 24 in the facility N20, and the recording medium 84n or the like is connected to the controller 86 of the image processing apparatus 44 of the facility X40. 44 statistical analysis processing units 58 can also be used.

  As described above, the CPU 72 executes the computer program of the present invention to achieve the object of the present invention. The computer program can be supplied to the computer CPU 72 in the form of a recording medium such as a CD-ROM 84n as described above, and a recording medium such as a CD-ROM 84n on which the computer program is recorded also constitutes the present invention. It will be. As the recording medium on which the computer program is recorded, for example, a memory card, a memory stick, a DVD, an optical disk, an FD, or the like can be used in addition to the recording medium described above.

  The brain image normal database 28 in Example 1 described above has a drawback that it is statistically very small, with 10 to 20 cases in categories such as age group. The third embodiment is characterized in that a statistical resampling method, particularly the bootstrap method described above, is applied to an artificial brain specimen. Statistics for the number of times can be obtained by hundreds to tens of thousands of arithmetic simulations by statistical resampling, and the above-mentioned difficulties can be solved by estimating parameter characteristics from the statistics. . As a result, it is possible to provide a brain image data processing system capable of improving the trial rate.

  In the third embodiment, as in the first embodiment, the standard deviation of each pixel of the brain image data recorded in the normal database 28 of the brain image is prepared in advance using the standard deviation calculation unit 38 of the facility N20. Here, the standard deviation is obtained by applying the bootstrap method as described above. The facility X40 can create an artificial brain coordinate sample of the facility X40 using the standard deviation.

  Next, an actual application example in which the average and the standard deviation are predicted from the small sample data of healthy persons in the facility N20 is shown. FIG. 8 shows an example of a prediction result by the bootstrap method by a frequency distribution (normal distribution). As shown in FIG. 8, the entire range is 0.82 to 1.18, and the display range is 0.85 to 1.15. Using the sample data of 20 cases, 20,000 trials were made by the bootstrap method.

  FIG. 9 shows an artificial brain coordinate obtained by fusing the average value of one example randomly extracted from the sample data of the above 20 examples and the standard deviation created from the data obtained as shown in FIG. The result of carrying out the statistical analysis process of a comparative case is shown. FIG. 9 shows a RENDER display similar to FIG. 3A, and each display portion shows a display result similar to FIG.

  FIG. 10 shows an example of a prediction result by the bootstrap method by a frequency distribution (Weibull distribution). As shown in FIG. 10, the entire range was 0.62 to 1.07, and the display range was 0.65 to 1.06. Using the sample data of 20 cases, 20,000 trials were made by the bootstrap method.

  FIG. 11 shows artificial brain coordinates obtained by fusing 10 samples randomly extracted from the sample data of the 20 cases and calculating an average value and standard deviations created from the data obtained as shown in FIG. The results of statistical analysis processing of comparative cases using are shown. FIG. 11 shows a RENDER display similar to that in FIG. 3A, and each display portion shows the same display result as in FIG.

  FIG. 12 shows an example of a prediction result by the bootstrap method using 10 sample samples randomly extracted from the sample data of the 20 examples, and a frequency distribution (gamma distribution). As shown in FIG. 12, the entire range was 0.71 to 1.12 and the display range was 0.78 to 1.10. Using the sample data of the above ten cases, 20,000 trials were made by the bootstrap method.

  FIG. 13 shows the result of statistical analysis processing of a comparative case using an artificial brain coordinate sample using the average value and standard deviation created from the data obtained as shown in FIG. FIG. 13 shows a RENDER display similar to that in FIG. 3A, and each display portion shows the same display result as in FIG.

  As described above, according to the third embodiment of the present invention, as in the first embodiment, each pixel of the brain image data recorded in the normal database 28 of the brain image using the standard deviation calculation unit 38 of the facility N20. The standard deviation is prepared in advance. Here, the standard deviation is obtained by applying the bootstrap method as described above. The facility X40 can create an artificial brain coordinate sample of the facility X40 using the standard deviation. As a result, in the facility X40, not only the healthy subject brain image database as the brain image database 48, but also the brain image database using the specialty specific to the clinical facility in the other facility N20 can be created. The third embodiment can be applied to creation of a brain coordinate sample for each disease by constructing a brain image database with few cases such as children and rare degenerative diseases that have been difficult to create.

  As an example of use of the present invention, in preparation of an artificial brain specimen and use of a normal database in a clinical facility where it is difficult to collect a large number of brain image data of healthy persons and patients or a clinical facility where it is difficult to use an expensive brain phantom Can be applied. Furthermore, it is possible to create a brain image database using the specialties peculiar to clinical facilities in other facilities N20. By constructing a brain image database with few cases such as children and rare degenerative diseases that have been difficult to create, it can be applied to the creation of brain coordinate samples by disease.

It is a figure which shows the image processing system of this invention. It is a flowchart which shows the flow of the brain image data processing program or method of this invention. It is a figure which shows the result of the "same condition" which performed the statistical analysis process for the test subject of the facility X40 by using the healthy subject sample of the same facility X40 as a control group, that is, the correct statistical analysis process result. Statistical analysis processing analyzed without applying the brain image data processing method of the present invention as a result of the “different conditions” obtained by subjecting subjects in the facility X40 to statistical analysis processing using healthy subjects from different facilities N20 as a control group It is a figure which shows a result. The result of “fusion conditions” analyzed by an artificial brain specimen obtained by fusing the standard deviation of the healthy subject specimen at facility N20 to the data (average value) of any healthy person A at facility X40, that is, the brain image data processing method of the present invention, etc. It is a figure which shows the statistical analysis process result analyzed by applying. The result of “fusion conditions” analyzed by an artificial brain specimen obtained by fusing the standard deviation of the healthy subject specimen at facility N20 to the data (average value) of any healthy person B at facility X40, that is, the brain image data processing method of the present invention, etc. It is a figure which shows the statistical analysis process result analyzed by applying. It is a block diagram which shows the internal circuit 70 of computers, such as the image processing apparatus 44 which performs the computer program of this invention. It is a frequency distribution (normal distribution) figure which shows the example of the prediction result by a bootstrap method. Statistical analysis of comparative cases using artificial brain coordinates that fuse the average value of one sample randomly extracted from the sample data of 20 cases and the standard deviation created from the data obtained as shown in FIG. It is a figure which shows the result of processing. It is a frequency distribution (Weibull distribution) figure which shows the example of the prediction result by a bootstrap method. Using artificial brain coordinates that fused the data obtained by randomly extracting 10 cases from the sample data of the 20 cases and calculating the average value and the standard deviation created from the data obtained as shown in FIG. It is a figure which shows the result of having performed the statistical analysis process of the comparative case. It is a frequency distribution (gamma distribution) figure which shows the example of the prediction result by the bootstrap method which extracted 10 examples from the sample data of the said 20 examples at random, and used the sample data of the 10 examples. It is a figure which shows the result of having performed the statistical analysis process of the comparative case using the artificial brain coordinate which used the average value and standard deviation which were created from the data obtained as FIG. 12 was used as it is.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Image processing system, 20 facilities N, 24,44 Image processing apparatus, 26,46 Imaging equipment, 28 Normal database of brain image, 30,50 program (block diagram), 32 1st normalization part, 34 1st brain image Data recording unit, 36 first average value calculating unit, 38 standard deviation calculating unit, 40 facility X, 45 display unit, 48 brain image database, 52 second normalization unit, 54 second brain image data recording unit, 56 second Average value calculation unit, 58 Statistical analysis processing unit, 60 Display processing unit, 70 Internal circuit, 72 CPU, 74 ROM, 76 RAM, 78 VRAM, 80 Image control unit, 82 Bus, 84a Disc, 84n CD-ROM, 86 Controller, 88 input operation unit, 90 input control unit, 92 external I / F unit.

Claims (10)

A brain image data processing system for causing a computer to execute processing of brain image data acquired by different imaging devices in different facilities,
The computer at the first facility
First normalization means for performing anatomical normalization using a stereotaxic brain coordinate system on brain image data acquired by a first imaging device under a first condition for a plurality of healthy persons in the facility;
First brain image data recording means for recording brain image data that has been subjected to anatomical normalization by the first normalization means in a healthy person brain image database of a first facility;
First average value calculating means for calculating an average value for each pixel or voxel of brain image data recorded in the brain image database of healthy persons in the first facility;
Standard deviation calculating means for obtaining a standard deviation for each pixel or voxel of the brain image data recorded in the brain image database of the healthy person in the first facility, using the average value obtained by the first average value calculating means. And
The computer at the second facility
A stereotaxic coordinate system was used for brain image data acquired by a second imaging device different from the first imaging device under a second condition different from the first condition for at least one healthy person in the facility. A second normalization means for performing anatomical normalization;
Second brain image data recording means for recording the brain image data that has been subjected to anatomical normalization by the second normalization means in a healthy person brain image database of a second facility;
When there are a plurality of healthy persons in the second normalization means, an average value is obtained for each pixel or voxel of brain image data recorded in the brain image database of healthy persons in the second facility, and the second normalization means When the number of healthy persons is one, the second average value calculating means for taking the average value of each pixel or voxel of the brain image data of the one person recorded in the healthy person brain image database of the second facility When,
Artificial brain image data having the average value obtained by the second average value calculating means and the standard deviation obtained by the standard deviation calculating means, and obtained by the second imaging device under the second condition for the subject Statistical analysis processing means for performing statistical analysis processing for comparing brain image data obtained by performing anatomical normalization using a stereotaxic coordinate system for the brain image data;
A brain image data processing system comprising: display processing means for displaying the statistical analysis processing performed by the statistical analysis processing means on a display unit.   2. The brain image data processing system according to claim 1, wherein the standard deviation calculation means uses a brain image data for each pixel or voxel recorded in the brain image database of healthy persons in the first facility to perform statistical analysis. A brain image data processing system characterized by obtaining a standard deviation for each pixel or voxel of brain image data by a sampling method.   3. The brain image data processing system according to claim 2, wherein the statistical resampling method in the standard deviation calculating means is a bootstrap method. A brain image data processing method that causes a computer to execute processing of brain image data acquired by different imaging devices in different facilities,
The computer at the first facility
A first normalization step for performing anatomical normalization using a stereotaxic brain coordinate system on brain image data acquired by the first imaging device under a first condition for a plurality of healthy persons in the facility;
A first brain image data recording step of recording the brain image data subjected to anatomical normalization in the first normalization step in a healthy person brain image database of a first facility;
A first average value calculating step for obtaining an average value for each pixel or voxel of the brain image data recorded in the brain image database of healthy persons in the first facility;
A standard deviation calculating step for obtaining a standard deviation for each pixel or voxel of the brain image data recorded in the brain image database of the healthy person in the first facility using the average value obtained in the first average value calculating step. When,
The computer at the second facility
A stereotaxic coordinate system was used for brain image data acquired by a second imaging device different from the first imaging device under a second condition different from the first condition for at least one healthy person in the facility. A second normalization step for performing anatomical normalization;
A second brain image data recording step of recording the brain image data subjected to anatomical normalization in the second normalization step in a healthy person brain image database of a second facility;
When there are a plurality of healthy persons in the second normalization step, an average value for each pixel or voxel of the brain image data recorded in the brain image database of healthy persons in the second facility is obtained, and the second normalization step A second average value calculating step in which the average value is a value for each pixel or voxel of the brain image data of the one person recorded in the brain image database of the healthy person in the second facility when there is one healthy person in When,
Artificial brain image data having the average value obtained in the second average value calculating step and the standard deviation obtained in the standard deviation calculating step, and acquired by the second imaging device on the subject under the second condition A statistical analysis processing step for performing a statistical analysis process for comparing the brain image data with the brain image data subjected to anatomical normalization using the stereotaxic brain coordinate system for the brain image data;
A brain image data processing method comprising: a display processing step for displaying the statistical analysis processing performed in the statistical analysis processing step on a display unit.   5. The brain image data processing method according to claim 4, wherein the step of calculating the standard deviation uses a brain image data for each pixel or voxel recorded in the brain image database of healthy persons in the first facility. A brain image data processing method characterized by obtaining a standard deviation for each pixel or voxel of brain image data by a sampling method.   6. The brain image data processing method according to claim 5, wherein the statistical resampling method in the standard deviation calculating step is a bootstrap method. A brain image data processing program that causes a computer to execute processing of brain image data acquired by different imaging devices in different facilities, the computer comprising:
Normalization means for performing anatomical normalization using a stereotaxic brain coordinate system on brain image data acquired by a predetermined imaging device under predetermined conditions for at least one healthy person in a facility;
Brain image data recording means for recording brain image data that has been subjected to anatomical normalization by the normalization means in a brain image database of healthy persons in the facility,
When there are a plurality of healthy persons in the normalizing means, an average value is obtained for each pixel or voxel of brain image data recorded in the healthy person brain image database of the facility, and there is one healthy person in the normalizing means. In this case, an average value calculating means that takes an average value of each pixel or voxel of the brain image data of the one person recorded in the brain image database of healthy persons in the facility,
Using the average value obtained by the average value calculating means and the computer of another facility different from the facility, the predetermined imaging under a condition different from the predetermined condition for a plurality of healthy persons in the other facility The brain image data obtained by another imaging device different from the device is subjected to anatomical normalization using a stereotaxic coordinate system, and the anatomical normalization is performed on the brain image data. Brain image data recorded in the brain image database of the healthy person in the other facility by obtaining the average value for each pixel or voxel of the brain image data recorded in the brain image database of the healthy person in the facility and using the average value Artificial brain image data having a standard deviation determined for each pixel or voxel of the above and brain image data obtained by the predetermined imaging device for the subject under the predetermined conditions Statistical analysis means for performing a statistical analysis to compare the anatomical normalize brain image data subjected with much brain coordinate system,
A program for causing a statistical analysis process performed by the statistical analysis processing means to function as a display processing means for displaying on a display unit.   8. The program according to claim 7, wherein the standard deviation calculating means uses the brain image data for each pixel or voxel recorded in the brain image database of the healthy person in the first facility, by a statistical resampling method. A program for obtaining a standard deviation for each pixel or voxel of image data.   9. The program according to claim 8, wherein the statistical resampling method in the standard deviation calculating means is a bootstrap method. A computer-readable recording medium on which the program according to claim 7 is recorded.

Images