JPH0998957A - Activated region extraction method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an activated region extraction method with no erroneous extraction of vein and the like regardless of effect of SN ratio fluctuation by a combined use of methods to use the t-test and a correlation coefficient, which are apt to be effected by the effect of SN ratio fluctuation, and a method to use a signal fluctuation ratio, which is not apt to be effected by the effect of SN ratio fluctuation. SOLUTION: At the first process, methods, which are apt to be effected by the effect of SN ratio fluctuation, are used to turn out function image data and to detect activated region. In this process, misjudgment of vein as an activated region can be avoided. In the second process, additional activated regions are detected by means of a method, which is not apt to be effected by the effect of SN ratio fluctuation, is used to turn not image data of signal fluctuation ratio as a function image data.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to brain function measurement MRI (hereinafter, fMR) using magnetic resonance imaging (hereinafter, MRI).
It belongs to the technical field relating to the data processing method in I). More specifically, the present invention belongs to the technical field related to a method for accurately extracting the area of the cerebral cortex activated by stimulation even when the SN ratio is low.

[0002]

2. Description of the Related Art fM for analyzing brain function using MRI
A technique called RI has been developed. With this technology,
The response in the cerebral cortex due to the application of stimuli such as light and sound
(Nuclear magnetic resonance) It is captured as a change in signal and imaged.
That is, an image taken while applying a stimulus to the subject (hereinafter,
A difference image is created from a stimulus-applied image) and an image captured without applying a stimulus (hereinafter, a rest image). A region with a high signal intensity on the difference image is regarded as a region that responds to the stimulus (hereinafter referred to as an activation region), and the functional area of the brain is identified and the rate of signal change is evaluated. Here, in this specification,
Image data created by using a rest image and a stimulus application image to extract an activation region like a difference image is called functional image data. The other image data is called morphological image data.

In order to accurately identify the functional area,
The method of extracting the activation region becomes particularly important. As a report comparing various extraction methods, Book of Abstrac
t, 12th Annual Meeting of
Society of Magnificent Reson
ance in Medicine, vol. 1,44
9 (1993) is known. In this report, the extraction method of the activation region is based on the signal intensity before the stimulation is applied,
The method is classified into two types: a method of evaluating and extracting a signal change rate associated with the application of a stimulus, and a method of evaluating and extracting a correlation between application of a stimulus and a signal change. Although it partially overlaps with this classification, the following three types of extraction methods are mainly used in practice. The first method is an extraction method that uses a signal change rate associated with the application of a stimulus. This method has a short data processing time when extracting the activation area, but has a drawback that not only the cerebral cortex but also veins are erroneously extracted. The second method is an extraction method that uses statistical processing such as t-test. This method has a feature that the ratio of the signal change rate to noise can be objectively evaluated. The third method is an extraction method using a correlation function. This method is an extraction method that evaluates the degree of synchronization between the applied stimulus and the signal change. Compared with the first method, the second and third methods have a feature that the data processing time at the time of extracting the activation region is longer, but the erroneous extraction of veins is less. Therefore, in many cases, the second or third method is used to extract the activation region.

In addition to the extraction method, the spatial resolution of the image is important in identifying the functional area. In order to accurately perform the identification, it is essential to improve the spatial resolution of the image, but this causes one problem. It is a reduction in the SN ratio per pixel. The signal amount of one pixel can be considered as the number of hydrogen atoms contained in one pixel. When the spatial resolution is improved, the volume of the real space corresponding to one pixel becomes smaller. As a result, the signal amount of one pixel is reduced and the SN ratio is reduced. The impact of the decrease in SN ratio is Book o
f Abstract, 2nd Annual Mee
toning of Society of Magneti
c Resonance, vol. 2,642 (199
4). That is, the area of the extracted activation region is reduced due to the following reasons. The area of the cerebral cortex in which the signal changes due to the application of the stimulus and the rate of change are constant regardless of the spatial resolution. Here, when the second or third method is used as the extraction method, if the SN ratio decreases due to the improvement of the spatial resolution, the area with a small signal change rate is not extracted as the activation area. for that reason,
The area of the extracted active region is reduced.

[0005]

As described above,
In the conventional activation area extraction method, many veins are erroneously extracted, and S
There is a problem that the area of the activation region changes due to the change of the N ratio. The object of the present invention is to maintain the characteristics of the second and third extraction methods described above, to prevent erroneous extraction of veins, etc.
It is to provide a method for extracting an activated region that is less susceptible to changes in the ratio.

[0006]

The activation region extraction method of the present invention is not easily influenced by the change of the SN ratio and the extraction method using the t-test or the correlation coefficient which is easily influenced by the change of the SN ratio. It is characterized by using the extraction method using the signal change rate together. In the present invention, the extraction method and S
The combined use of the extraction method that is less likely to be affected by the N ratio makes it possible to satisfactorily extract the activation region even when the SN ratio decreases as the spatial resolution improves.

[0007]

Embodiments of the present invention will be described below in detail with reference to the drawings. First, FIG. 2 shows an example of measurement of fMRI, which is one of the fields to which the present invention can be applied. fMR
In I, a stimulus application period is provided during measurement, and images are taken at regular time intervals. In this time-series image, the process of evaluating the time change of the signal for each coordinate and extracting the activation region is the activation region extraction method. When extracting the activation region, one extraction method is usually used, but in the present invention, two kinds of extraction methods are used. That is, the functional image data created by using the extraction method that is easily influenced by the SN ratio of the image and the functional image data created by using the extraction method that is not easily influenced by the SN ratio are used together for extraction. First, two types of extraction methods will be described, and then the extraction procedure will be described.

The extraction method which is easily influenced by the SN ratio means that 1) there is a term for inputting signal variance or deviation in the mathematical expression expressing the extraction method, and 2) the value becomes large when the noise becomes large. Is an extraction method corresponding to any of the following. Note that the noise here is the time variation of the signal at the same coordinates of the time series image, and the signal variance is the deviation from the time average value of the signals at the same coordinates of the time series image. An example of 1) is an extraction method using a t-test. (Equation 1) shows a calculation formula of t value.

[0009]

[Formula 1] t = (X ₁ −X ₂ ) / {√ {(D ₁ + D ₂ ) / (n ₁ + n ₂ −2)} · √ {1 / n ₁ + 1 / n ₂ }} (Formula 1 Here, X is the average value of the signal, D ₁ and D ₂ are the sum of squares of the deviation of the signal, and n is the number of images. In (Equation 1), √ {}
Is the square root within {}.

The subscripts 1 and 2 represent the rest period and the stimulus application period, respectively. As can be seen from (Equation 1),
When the value of D ₁ and D ₂ is increased, the value of t is reduced. Since the activation region is extracted by setting a threshold value for the value of t, it can be said that the extraction method using the t-test is easily affected by the SN ratio.
As an example of 2), there is an extraction method using a cross-correlation function, and (Formula 2) shows an example thereof.

[0011]

## EQU2 ## f (S, E) = SE / {| S ||||||||} | In (Equation 2), ‖S‖ and ‖E‖ represent the absolute values of the vectors S and E, respectively. E is
It is a user vector that can be arbitrarily set by the user. The user vector has a value of -1 during the rest period and 1 during the stimulation application period. While the signal vector is different for each coordinate, the user vector uses the same vector for each coordinate. (Equation 2) is an inner product (SE) of the signal vector S and the user vector E, where the number of time-series images is the number of dimensions. Therefore, the smaller the angle formed by both vectors, the larger the value of (Equation 2). When a correlation function is used to extract the activation area, a threshold value is set for this value for extraction. Now, although there is no term for inputting the signal variance in (Equation 2), as the noise increases, the absolute value of the signal value vector S increases, and as a result, the value of the cross-correlation function decreases. To do. This is shown below.
However, for simplification, the rest period and the stimulus applying period will be described without distinction. When the number of images is sufficiently large, the sum of the noises N _i becomes 0 as shown in (Equation 3), and it is assumed that the average value S _{m of the} signals does not change even if the SN ratio changes. In (Equation 3), the subscript i represents the number of the time-series image, and the addition Σ is performed for i = 1, 2, ..., N.

[0012]

ΣN _i = 0 (Equation 3) The signal value S _i of each image is considered to be the sum of the average value X _{m of the} signal and the noise N _i . Here, a coefficient a representing the magnitude of noise
And b (a <b) are used, the signal S _ai with small noise and the signal S _bi with large noise are respectively expressed by ( _Equation 4),
It can be expressed as (Equation 5).

[0013]

[Equation 4] S _ai = S _m + aN _i (Equation 4)

[0014]

S _bi = S _m + bN _i (Equation 5) Since the same user vector is used regardless of the change in the SN ratio, the denominator of (Equation 2) (the absolute value of the signal value vector S and the user vector E is The magnitude of the product of the values depends on the magnitude of the signal value vector S. Therefore, the change in the magnitude of the signal value vector with respect to the change in the SN ratio is calculated by (Equation 4) and (Equation 5)
Are compared using. (Equation 6) is the square of the magnitude of the signal value vector when the noise is small. Also,
(Equation 7) is the square of the magnitude of the signal value vector when noise is large. It can be seen that the difference between (Equation 6) and (Equation 7) is (Equation 8), which is always a positive value. That is,
There is a term whose value increases whenever the signal variance increases. In addition, in (Equation 6), (Equation 7), (Equation 8),
The addition Σ is performed for i = 1, 2, ..., N. (Equation 8) shows the difference between the squared value of the signal value when the noise is large and the squared value of the signal value when the noise is small.

[0015]

Σ {S _ai } ² = Σ (S _m + aN _i ) ² = ΣS _m ² + a ² ΣN _i ² + 2aS _m ΣN _i = ΣS _m ² + a ² ΣN _i ² (Equation 6)

[0016]

Σ {S _bi } ² = Σ (S _m + bN _i ) ² = ΣS _m ² + b ² ΣN _i ² + 2bS _m ΣN _i = ΣS _m ² + b ² ΣN _i ² (Equation 7)

[0017]

Σ {S _bi } ² −Σ {S _ai } ² = (b ² −a ² ) ΣN _i ² (Equation 8) As described above, the extraction method using the t test and the correlation function are used. It can be judged that the extracted method is susceptible to the change of the SN ratio. An autocorrelation function may be used to extract the activation area.

On the other hand, examples of the extraction method less susceptible to the SN ratio include an extraction method using a signal change rate. This is because the averaging process is performed when the signal change rate is calculated, so that the influence caused by the change in the SN ratio is reduced.

Next, the procedure for extracting the activated area using these extraction methods will be described.

The method of extracting the activated region using the present invention is 2
It is divided into stages. In the first stage, the activation region is extracted without erroneous extraction, and in the second stage, the region extracted in the first stage is used as a starting point to expand the activation region. Various area expansion methods have been proposed in the field of image processing, and these methods can also be used in the present invention. However, these methods are different in that the user determines and specifies the starting point of the region expanding process from the structure of the target object, but the present invention extracts the starting point in the first process. Further, the information of the functional image data created by the extraction method used in the first step and the information of the functional image data created by the extraction method used in the second step are different, and this different information The point that the area is expanded by using both is also different from the conventional area expansion processing.

FIG. 1 is a flow chart showing an example of a data processing procedure of the activation area extraction method using the present invention. In the first process (process 1), functional image data (functional image data susceptible to the influence of the SN ratio) is created using an extraction method susceptible to the SN ratio, and the activation region is extracted. These extraction methods can extract the activation region without erroneously extracting the vein. Second process (second process)
Then, the area that has not been extracted as the activated area in the first process is re-evaluated. For this evaluation, signal change rate image data, which is functional image data created using an extraction method that is not easily influenced by the SN ratio, is used. First, the activation region extracted in the first process is set as the starting point, and the signal change rate of the pixel adjacent to the starting point is evaluated. That is, if the signal change rate is equal to or higher than a predetermined threshold value, the active area is expanded. Next, the signal change rate of the pixels in the vicinity of the activated area after the expansion is evaluated, and if the signal change rate is equal to or higher than a predetermined threshold value, the pixel is similarly expanded as the activated area. The expansion of the activation region is performed until no adjacent pixels satisfy the threshold value condition of the signal change rate.

FIG. 4 shows this state. When expanding the activation region, not only the signal change rate but also the spatial gradient of the signal change rate may be used. By performing a series of processing related to the area expansion, it is possible to prevent the activation area from being reduced due to the change of the SN ratio. On the other hand, in order to prevent the activation region from overflowing outside the cerebral cortex, a threshold may be set for the signal value of the time-series image to extract the cerebral cortex, and the region to be searched for the activation region may be determined. In this case, two types of functional image information and morphological image information are used together.

In the first processing and the second processing,
It is necessary to set a threshold value for each. The present invention is not limited to the setting method of each threshold value. That is,
The threshold value may be changed each time the activation region is extracted, or the preset threshold value may be repeatedly used for extraction. In the description of the above embodiment, M
Although only the measurement using RI has been described, the application target of the present invention is not limited to MRI.

FIG. 5 is a schematic configuration diagram of a magnetic resonance apparatus to which the method of the present invention is applied. In FIG. 5, 621 is a magnet that generates a static magnetic field H ₀ , 622 is a measurement target, 623 is a coil for generating a high frequency magnetic field and detecting a magnetic resonance signal generated from the measurement target 622 ,. 6624, 625, 62
Reference numeral 6 is a gradient magnetic field generating coil for generating gradient magnetic fields in the x-axis direction, the y-axis direction and the z-axis direction, respectively. 627, 6
28 and 629 are the gradient magnetic field generating coils 624 and 62.
5, a gradient magnetic field driving device for supplying a current to 626.
The gradient magnetic field generating coil and the gradient magnetic field driving device are collectively referred to as a gradient magnetic field system. Reference numeral 630 is a calculator for calculating the measured data, and 631 is a display for displaying the calculation result of the calculator 630.

Next, an outline of the operation of this inspection apparatus will be described.
High-frequency magnetic field H for exciting the nuclear spins of the measurement target 622
_{In the} modulator ₁ , the high frequency generated by the synthesizer 632 is waveform-shaped and power-amplified by the modulator 633, and a current is supplied to the coil 623 to generate it. Coil driving device 627, 6
The gradient magnetic field generating coils 624, 625, 626 supplied with current from 28, 629 generate a gradient magnetic field and modulate the magnetic resonance signal from the measurement target 622. This modulated signal is received by the coil 623, amplified by the amplifier 634, detected by the detector 635, AD converted, and then calculated by the computer 6.
30 is input. After the calculation, the computer 630 displays the calculation result on the display 631. It should be noted that the computer 630 controls each device so that it operates at a preprogrammed timing and intensity. Among these programs, a program that describes the high frequency magnetic field, the gradient magnetic field, the timing and intensity of signal reception, is called a pulse sequence.

In summary, the present invention is summarized as follows: A stimulus is applied to a subject for a predetermined period of time to capture time-series images, and the signal intensity at the same coordinates of the time-series images is input to a predetermined determination formula. In the activation area extraction method that creates functional image data with the result as a value and extracts the active area in the brain (hereinafter referred to as the activation area) using the functional image data, it is superimposed on the signal and stimulated. The process of creating functional image data that reflects changes in the SN ratio and extracting the activation area by using a judgment formula that includes a term whose value always increases as the uncorrelated fluctuation (hereinafter, noise) increases. The activation area extraction is characterized in that it includes a process of creating functional image data that does not reflect changes in the SN ratio and extracting the activation area by using a judgment formula that does not include a term whose value increases without exception. In law Further, the extraction method of the active region,
Using the process of setting the region extracted using the functional image data reflecting the change of the SN ratio as the start region of the activation region and the functional image data adjacent to the activation region and not reflecting the change of the SN ratio The area obtained by the area expansion processing is activated by including the import processing that merges the extracted area with the active area and the area expansion processing that repeats the import processing until there is no area newly included in the active area. The feature is that it is extracted as a digitized region.

[0027]

According to the present invention, the activated region can be accurately extracted even when the SN ratio is lowered due to the improvement of the spatial resolution.

[Brief description of drawings]

FIG. 1 is a flowchart showing an activated area extraction procedure using the present invention.

FIG. 2 is a diagram showing a measurement example in FMRI.

FIG. 3 is a diagram showing a signal vector.

FIG. 4 is a diagram illustrating a procedure of an activation region according to the present invention.

FIG. 5 is a diagram showing a configuration example of an MRI apparatus which is an example to which the present invention is applied.

[Explanation of symbols]

1 ... First processing in activated area extraction method using the present invention, 2 ... Second processing in activated area extraction method using the present invention
, 621 ... Magnet for static magnetic field generation, 622 ... Target for measurement, 623 ... Coil for high frequency magnetic field generation and signal detection,
624, 625, 626 ... Gradient magnetic field generating coil, 62
7, 628, 629 ... Coil drive device, 630 ... Calculator, 631 ... Display, 632 ... Synthesizer, 6
33 ... Modulator, 634 ... Amplifier, 635 ... Detector.

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] January 23, 1996

[Procedure Amendment 1]

[Document name to be amended] Drawing

[Correction target item name] All figures

[Correction method] Change

[Correction contents]

FIG.

[Figure 3]

[Fig. 2]

FIG. 4

[Figure 5]

Claims (11)

[Claims] 1. A stimulus is applied to a subject for a predetermined period of time to capture time-series images, and the signal intensity at the same coordinates of the time-series images is input to a predetermined determination formula, and the result of the determination formula is determined. In the extraction method of the active region, which creates the functional image data having a value, and extracts the active region that is an active region in the brain using the functional image data, the stimulus is superposed on the signal. The functional image data reflecting fluctuations in the SN ratio, which is a fluctuation having no correlation, and which uses the determination formula including a term in which the time variation of the signal at the same coordinates of the time-series image is always large when the signal is large. And the functional image data that does not reflect the change in the SN ratio, using the first processing for creating an active region and the determination formula that does not include a term whose value always increases as the time fluctuation of the signal increases. Creates the activity Activation region extraction method and having a second processing to extract the area. 2. The active area extraction method according to claim 1, wherein the area extracted by the first processing is set as a start area of the active area, and the area is adjacent to the active area, and A region expansion process of merging the region extracted by the second process with an activation region and repeating the merging process until there is no region newly merged with the activation region. An active area extraction method, characterized in that an area obtained by processing is extracted as the activation area. 3. The active area extraction method according to claim 2, further comprising a process of determining an area to search for the active area from a signal value of the time series image. 4. The active area extraction method according to claim 1, wherein the areas extracted by both the functional image data reflecting the change of the SN ratio and the functional image data not reflecting the change of the SN ratio are activated. The process of setting the start area of the area,
A region that is adjacent to the activation region and merges the region extracted by the second process with the activation region, and repeats the merge until there is no region newly merged with the activation region. An active area extraction method comprising: an expansion process, and extracting an area obtained by the area expansion process as the activation area. 5. The active area extraction method according to claim 4, further comprising a process of determining an area to search for the active area from a signal value of the time series image. 6. The activation region extraction method according to claim 1, wherein the determination formula used to create the functional image data reflecting the change in the SN ratio is a determination formula using a t-test. Characterized activation area extraction method. 7. The activation region extraction method according to claim 1, wherein the determination formula used to create the functional image data reflecting the change in the SN ratio is a determination formula using an F test. Characterized activation area extraction method. 8. The activation region extraction method according to claim 1, wherein the determination formula used to create the functional image data reflecting the change in the SN ratio is a determination formula using a chi-square test. A method for extracting activated regions characterized by. 9. The activation region extraction method according to claim 1, wherein the determination formula used to create the functional image data reflecting the change in the SN ratio is a determination formula using an autocorrelation function. A method for extracting activated regions characterized by. 10. The activation region extraction method according to claim 1, wherein the determination formula used to create the functional image data reflecting the change in the SN ratio is a determination formula using a cross-correlation function. A method for extracting activated regions characterized by. 11. The activation region extraction method according to claim 1, wherein the determination formula used for creating the functional image data that does not reflect the change in the SN ratio is the signal intensity during the period when the stimulus is applied. A method for extracting an activation region, characterized in that a signal change rate is calculated from a signal intensity during a period in which a stimulus is not applied, and the activation region is extracted by setting a threshold value on the signal change rate.