JPH09238914A - Brain function measurement data processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve accurate estimation of changes in the quantity of local blood stream and in the consumption of oxygen by a method wherein an optical path obtained by actual measurement or simulation is used to extract pixels existing along the optical path and then, changes in signals detected by optical brain function measurement are distributed among the pixels of an fMRI on the optical path. SOLUTION: Weight coefficients are allotted to pixels of an image of fMRI based on the positions of an irradiation end and a detection end and a probability model of an optical path distribution (processing 1). Then, an integrated value of signal values and the weight coefficients of the fMRI image is calculated per pixel (processing 2) to determine the sum of the integrated values (processing 3). Thereafter, changes in signals obtained by optical brain function measurement are divided by the integrated values previously calculated to calculate a correction factor C (processing 4). Then, the weight coefficients of the pixels, the signal values of the fMRI image and the correction factor are integrated (processing 5) to calculate changes in signals in the optical brain function measurement at the positions of the pixels of the fMRI image. The results are displayed to be superimposed on the MRI image (processing 6).

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to brain function measurement (hereinafter, fMRI) using a magnetic resonance imaging (hereinafter, MRI) apparatus, and
The present invention relates to brain function measurement using light (hereinafter, optical brain function measurement).

[0002]

2. Description of the Related Art The structure of an MRI apparatus is shown in FIG. 101
Is a magnet for generating a static magnetic field, 102 is an imaging target such as a subject, 103 is a bed on which the imaging target 102 is placed, 104 is a high-frequency magnetic field coil for generating an RF magnetic field and at the same time detecting an echo signal generated from the imaging target 102. ,
108, 109, and 110 are x-direction, y-direction, and z-direction, respectively.
It is a gradient magnetic field generation coil for generating a gradient magnetic field in a direction. Reference numerals 105, 106 and 107 denote coil driving devices for supplying currents to the gradient magnetic field generating coils 108, 109 and 110, respectively. Reference numeral 115 is a computer for processing the measured data and performing image reconstruction, and 116 is a CRT display for displaying the reconstructed image of the computer 115.

An outline of the operation of the device will be described. Shooting target 1
The high frequency magnetic field that excites the nuclear magnetization of 02 is generated by the synthesizer 1.
The high frequency generated by 11 is subjected to waveform shaping and power amplification by the modulator 112, and is generated by supplying a current to the high frequency magnetic field coil 104. The echo signal from the imaging target 102 is received by the high-frequency magnetic field coil 104,
After being amplified by the amplifier 113 and detected by the detector 114,
It is input to the computer 115 and stored in the memory 117. The memory 117 also stores data being processed and final results. The computer 115 performs image reconstruction and displays the result on the CRT display 116.

An example of the fMRI measurement method using this apparatus is shown in FIG. In this measurement, time series images are taken at regular time intervals. At that time, a period for applying a stimulus such as light or sound to the subject is provided. The signal intensities of the image before the stimulus application and the image during the stimulus application period are compared for each pixel, and it is determined that the pixel in which the signal intensity changes is activated by the stimulus (hereinafter, activated region). The signal change in the activation region is, for example, as shown in FIG.

The purpose of fMRI is to extract the activation region to grasp the functional distribution of the brain, and to observe the signal change rate in the activation region. In particular, it is expected that early diagnosis of brain diseases will be possible due to the large rate of signal change. However, a plurality of factors are mixed in the signal change in fMRI, and there is a problem that the value of the signal change rate does not reflect the nerve activity itself.

Below, the cause of signal change in fMRI and this problem will be explained. MR by applying stimulus
It is considered that the change in signal is caused by the difference in magnetic properties between oxyhemoglobin and deoxyhemoglobin and the change in blood flow. FIG. 6 is a flowchart showing a state until the MR signal is changed by applying a stimulus. First, application of a stimulus excites nerve cells in the cerebral cortex and consumes oxygen. This locally increases the amount of reduced hemoglobin. Next, since the oxygen consumption increases in the activation region, the blood flow of arterial blood increases. This increase in blood flow supplies the activated region with an amount of oxygen that far exceeds the amount consumed by the application of the stimulus. That is, in the activated region, a change in magnetic properties due to an increase in oxyhemoglobin and an increase in blood flow occur at the same time. Since both of these affect the MR signal, fM
It is considered that two factors are mixed in the RI signal change.

In order to clarify the cause of brain disease, it is necessary to separate the factors of signal change. For example, when the change in oxygen consumption is small, the response of the cranial nerve to the stimulus is small, so it can be determined that a disease has occurred in the cerebral cortex. On the other hand, when the change in blood flow is small, it can be determined that there is no abnormality in the cerebral cortex and that a disease has occurred in the vascular system. Thus, it is important to separate the respective factors and observe the signal change, but it is difficult to separate them by fMRI alone.

On the other hand, optical brain function measurement is available as a measurement method capable of independently measuring changes in the concentrations of reduced hemoglobin and oxyhemoglobin. Next, this measuring method will be described.

FIG. 7 shows a device configuration diagram of the optical brain function measurement.
The irradiating light is 0.5 kHz with the oscillator 201 of 0.5 kHz.
Each of them is intensity-modulated by the oscillator 202 of Hz and has a wavelength of 780 nm and a laser diode 203 of wavelength 780 nm.
It is output from the irradiation end 211 by using the laser diode 204 of m. The transmitted light collected at the detection end 212 is detected by the avalanche photodiode 205 and is synchronously detected by the two lock-in amplifiers 206 and 207. After that, A / D conversion is performed by the AD converter 208, and the result is stored in the computer 209.

The apparatus shown in FIG. 7 can measure the transmitted light intensity of two wavelengths independently. That is, it is possible to separate changes in blood flow from changes in oxygen consumption by calculating changes in the concentrations of oxyhemoglobin and deoxyhemoglobin from the measured transmitted light intensity of two wavelengths (FIG. 8). However, since the signal change obtained by optical brain function measurement is the sum of reactions that occur in the region existing on the path through which photons pass (hereinafter referred to as the optical path), the spatial resolution of optical brain function measurement is compared with fMRI. Is inferior.

[0011]

SUMMARY OF THE INVENTION The object of the present invention is to find fM
Accurate estimation of changes in local blood flow and oxygen consumption based on measurement results of RI and optical brain function measurement.

[0012]

In order to solve this problem, the present invention uses an optical path obtained by actual measurement or simulation to improve accuracy. First, in the time series image of fMRI, pixels existing along the optical path are extracted. Next, the signal change detected by the optical brain function measurement is distributed to the fMRI pixels existing on the optical path.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described with reference to the drawings. The fMRI images are 1) time series images,
2) Time-series functional images obtained by subjecting time-series images to normalization processing with signal values during the rest period, filter processing such as moving average, averaging processing, and statistical processing such as t-test, 3) In 2) above A threshold value is set for the processing result described above, and a binary image in which the signal value of a pixel satisfying the threshold condition is 1 and the signal value of the other pixels is 0 is included. These processes have no problem in applying the present invention.

FIG. 2 is a diagram in which the optical path 220 is superposed on a diagram in the vicinity of the activated region extracted by fMRI. Incident light is emitted from the irradiation end 211, and the pixels 300 to
After passing through 306, it is detected at the detection end 212. That is, the attenuation of the light intensity due to passing through the optical path 220 is
It reflects the hemodynamics due to neural activity in pixels 300 to 306. Therefore, by applying the processing shown in Expression 1 to the result obtained by fMRI, the difference in spatial resolution between fMRI and optical brain function measurement can be corrected. However, P in Equation 1 is a signal for optical brain function measurement, and S
Is the signal intensity of the fMRI image, C is the correction coefficient for fMRI and optical brain function measurement, and the subscript i is the pixel number in the fMRI image. The pixel numbers are n1 to n2. In addition, in the equation 1, the pixels 300 to 306 are targeted for signal processing, but the pixels corresponding to the positions of the scalp and the skull like the pixels 300 and 306 do not change in signal intensity even when a stimulus is applied. Therefore, you may exclude from the object of signal processing.

[0015]

[Equation 1]

The actual optical path through the brain is shown in FIG.
It is not as simple as the optical path 220 of FIG.
Like the optical path 221 in the frame of, there is a spread due to scattering. It is also known that the optical path passing through this spread has a certain probability distribution. For this probability distribution,
For example, Radar and Sonar Part II (Radarand
Sonar PartII), p57, Springer-Verlag (1990). Further, generally, this distribution is calculated by a simulation using a Monte Carlo method or a simulation based on a diffusion equation.

Therefore, in order to perform a more accurate comparison by the optical brain function measuring method, it is considered necessary to weight each pixel of the fMRI image using a predetermined probability model and process the data. The method of processing the measurement result of fMRI in this case is shown in Equation 2. Wi represents the weighting coefficient of the pixel i of the fMRI image.

[0018]

[Equation 2]

These equations 1 and 2 are processing methods in the case where there is one irradiation end and one detection end. However, in actual optical brain function measurement, since multiple terminals are attached to the head of the subject and switching is performed by switching the irradiation end and the detection end, the irradiation end and the detection end, and the weighting factor may change. become. This can be dealt with by expanding Formula 2 to Formula 3. However, in Equation 3, the subscript m represents the position of the irradiation terminal, and the subscript j represents the position of the detection terminal. That is, mjW
i represents the weighting coefficient of the pixel i of fMRI when the irradiation terminal is at the position m and the detection terminal is at the position j.

[0020]

(Equation 3)

By using the equation 3, even if the number of terminals increases, it is possible to easily process the measurement result of fMRI and compare it with the result of optical brain function measurement. The value of the correction coefficient C can be calculated because P, Wi, and Si are known amounts obtained from the respective measurements.

Up to now, the processing method for the result of fMRI has been described, but by obtaining the value of the correction coefficient C included in the equations 1 to 3, the position information is approximated to the signal change in the optical brain function measurement. Can be given. That is, data of optical brain function measurement is assigned to each pixel of the fMRI image, and it is possible to associate the fMRI signal change at a predetermined position with the hemoglobin amount change.

The correction coefficient C must be calculated in order to make this correspondence, but this is possible as described above. Equation 4 shows an equation for assigning the data of optical brain function measurement to each pixel of the MRI image. The signal Pi of the photobrain function measurement at the pixel number i of the fMRI image is calculated based on the weighting factors Wi and fMR.
It is calculated by the product of the signal intensity Si of the I image and the correction coefficient C.

[0024]

(Equation 4)

The processing method for the result of fMRI and the processing method for the result of optical brain function measurement have been described. The analysis procedure using these processing methods will be described below.

FIG. 1 is a diagram showing an example of an analysis procedure of the result of fMRI and the result of optical brain function measurement. However, fM
It is assumed that the measurement results of RI and optical brain function measurement, and the probabilistic model have already been obtained.

First, a weighting factor for each pixel of an fMRI image is assigned from the positions of the irradiation end and the detection end and the probability model of the optical path distribution (process 1). Next, the integrated value of the signal value of the fMRI image and the weighting coefficient is calculated for each pixel (process 2), and the total sum of the integrated values is obtained (process 3). This is the process represented by Formula 3. After that, the signal change obtained by the optical brain function measurement is divided by the sum of the integrated values calculated previously to calculate the correction coefficient C (process 4). Weighting factor of each pixel and fMR
The signal value of the I image and the correction coefficient C are integrated (process 5), and the signal change of the photobrain function measurement at the position of each pixel of the fMRI image is calculated. This is the process represented by Equation 4. Further, if necessary, the result of the process 5 may be displayed in an overlapping manner on the MRI image (process 6).

Next, an example of selecting an fMRI image to be processed 3 will be described. As shown in FIG. 9A, the process 3 is basically performed on one of the fMRI images existing in time series. However, this process is expanded to include a plurality of fM images.
It is also possible to target the RI image. Note that process 3
Hatching is applied to the fMRI image to be processed. When the same measurement is repeated many times, the fMRI image of the same time phase of each session is processed 3 as shown in FIG. 9B.
May be the target of. In the process of FIG. 9B, the signal change of each session is averaged, so that there is an effect of reducing the signal variation.

FIG. 9 shows a process 3 for fMRI images of the same time phase.
However, it is also possible to apply the process 3 to the fMRI images of a plurality of time phases. Such a processing method is preferably applied when the time-series image capturing interval of fMRI is superior to the time resolution of optical brain function measurement.

FIG. 10A is a diagram showing an example of application of the process 3 to the fMRI images of a plurality of time phases. This also has the effect of averaging the signal changes and reducing signal variations. Furthermore, when the same measurement is repeated many times, FIG.
As described above, the fMRI images in a plurality of time phases of each session may be the target of the process 3.

[0031]

According to the present invention, it is possible to accurately estimate changes in local blood flow and oxygen consumption from the measurement results of fMRI and optical brain function measurement.

[Brief description of drawings]

FIG. 1 is a flowchart showing an analysis procedure by fMRI and optical brain function measurement according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram in which an optical path is superimposed on an fMRI image.

FIG. 3 is an explanatory diagram showing an optical path passing through the brain (a) when ignoring scattering and an optical path when (b) considering scattering.

FIG. 4 is a block diagram of an MRI apparatus.

FIG. 5 is an explanatory diagram showing an example of (a) measurement method in fMRI and (b) signal change in an activated region.

FIG. 6 is a flowchart showing the steps until the MR signal is changed by applying a stimulus.

FIG. 7 is an explanatory diagram of a brain function measuring device using light.

FIG. 8 is a characteristic diagram of signal changes in brain function measurement using light.

FIG. 9 is an explanatory diagram showing an application example of (a) an fMRI image of one session and (b) an application example of an fMRI image of a plurality of sessions to fMRI images of the same time phase.

10A and 10B are explanatory views showing an example of application of process 3 to an fMRI image of a plurality of time phases, and (a) an example of application to an fMRI image of the same session;

[Explanation of symbols]

1 ... Weighting coefficient assignment processing, 2 ... Integration processing of signal value of fMRI image and weighting coefficient, 3 ... Addition processing of integrated value obtained in processing 2, 4 ... Correction coefficient calculation processing, 5 ... Optical brain function measurement Signal change allocation processing in 6 ... Mapping processing of signal change in optical brain function measurement.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kenichi Okajima 1-280, Higashikoigokubo, Kokubunji, Tokyo Inside the Central Research Laboratory, Hitachi, Ltd.

Claims (17)

[Claims] 1. A static magnetic field generating means for generating a static magnetic field having a spatially uniform intensity, and a gradient magnetic field generating for generating a gradient magnetic field having an intensity gradient in each of X, Y and Z directions orthogonal to each other. Means, a high-frequency magnetic field generating means for generating a high-frequency magnetic field for exciting the nuclear magnetization of the subject, a detecting means for detecting a nuclear magnetic resonance signal from the subject, a computer for calculating the detection signal of the signal detecting means, Using an MRI apparatus having a calculation result output unit by a computer, measurement is performed by providing a stimulus application period during which a subject's cranial nerves are excited during a time-series image capturing period, and then performing predetermined processing on the time-series image. 300 nm of brain function MRI measurement results in which time-series functional images were created and the site that responded to the stimulus was extracted
From the light generating means for generating light having a wavelength of 1300 nm, a light irradiating means for irradiating the subject with the light generated by the light generating means, a light detecting means for detecting light transmitted through the subject, and the light detecting means. Using the optical measurement device that has a calculator that calculates the signal from the, and using the measurement result of the optical brain function measurement that observes the change of the detection signal by providing the stimulus application period during the measurement period, and passes through the body of the subject Brain function measurement data processing characterized by setting an optical path to be performed on an image, extracting a pixel of a time-series functional image included in the optical path, and performing arithmetic processing using a signal value of the extracted pixel Law. 2. The brain function measurement data processing method according to claim 1, wherein the optical path is described by a function of a probability distribution in consideration of light scattering. 3. The brain function measurement data processing method according to claim 2, wherein the function of the probability distribution is calculated by a simulation simulating light scattering. 4. The brain function measurement data processing method according to claim 3, wherein the simulation simulating the light scattering is a simulation using a Monte Carlo method. 5. The brain function measurement data processing method according to claim 3, wherein the simulation simulating the light scattering is a simulation based on a diffusion equation. 6. The brain function measurement data processing method according to claim 1, wherein the calculation process includes a process of calculating a sum of signal values of the extracted pixels. 7. The brain function measurement data processing method according to claim 6, wherein the process of calculating the sum of the signal values of the extracted pixels is performed for one time-series function image in one process. 8. The brain function measurement data processing method according to claim 6, wherein the process of calculating the sum of the signal values of the extracted pixels is performed for a time-series functional image of a temporary phase in one process. 9. The brain function measurement data processing method according to claim 6, wherein the process of calculating the sum of the signal values of the extracted pixels is performed for a time series functional image of a plurality of time phases in one process. . 10. The brain function measurement data processing method according to claim 1, wherein the calculation process includes a process of setting a weighting coefficient for each pixel of the time-series function image using a function of the probability distribution of the optical path. 11. The brain function measurement data processing method according to claim 1, wherein the calculation process includes a process of calculating a product of a weighting coefficient of each pixel and a signal value and calculating a sum of the product values. 12. The brain function measurement data processing method according to claim 11, wherein the process of calculating the sum of the product values is performed for one time-series function image in one process. 13. The brain function measurement data processing method according to claim 11, wherein the process of calculating the sum of the product values is performed for a time-series functional image of a temporary phase in one process. 14. The brain function measurement data processing method according to claim 11, wherein the process of calculating the sum of the product values is performed for a time series functional image of a plurality of time phases in one process. 15. The arithmetic processing standardizes the measurement result of optical brain function measurement using the signal sum calculated in any one of claims 6 to 9 and claims 11 to 14, and converts both function measurement results. The brain function measurement data processing method according to claim 1, which is a process for obtaining a coefficient. 16. The brain according to claim 1, including a distribution process of calculating the product of the weighting factor, the signal value of the time-series functional image, and the conversion coefficient, and allocating the calculation result to each pixel of the time-series functional image. Functional measurement data processing method. 17. The distribution process according to claim 1, wherein the product of the weighting factor, the signal value of the time-series functional image, and the transform coefficient is calculated, and the calculation result is assigned to each pixel of the time-series functional image. A brain function measurement data processing method including a mapping process of optical brain function measurement, in which the result of the distribution process is superimposed and displayed on an image taken by an MRI apparatus.