JP3387236B2 - Biomagnetic measurement device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a biomagnetism measuring apparatus for estimating the position, direction, and size of a living activity current source, and more particularly to an estimated living activity. The present invention relates to a technique for accurately evaluating the position, orientation, and size of a current source. 2. Description of the Related Art When a living body is stimulated, polarization formed across cell membranes is broken, and a living activity current flows.
This biological activity current appears in the brain, heart and the like, and is recorded as an electroencephalogram and an electrocardiogram. The magnetic field generated by the biological activity current is recorded as a magnetoencephalogram or a magnetocardiogram. In recent years, as a device for measuring a minute magnetic field in a living body, a SQUID (Super-conducting Quantum Inter
face Device: A sensor using a superconducting quantum interferometer) has been developed. In this SQUID sensor, a magnetic sensor is housed in a container called a dewar, immersed in a refrigerant such as liquid nitrogen. The SQUID sensor includes a single-channel SQUID sensor and a multi-channel SQUID sensor.
The ID sensor includes one magnetic sensor, and the multichannel SQUID sensor includes a plurality of magnetic sensors. [0004] For example, when a single-channel SQUID sensor is used to measure a minute magnetic field based on a biological activity current in the brain, the SQUID sensor is moved to a plurality of measurement positions near the subject's head. Magnetic field data at each measurement position is detected and collected in a non-invasive manner to obtain a plurality of measurement data. In addition, the same measurement is performed using a multi-channel SQUID.
When using the D sensor, the SQUID sensor is placed close to the subject's head (at this time, each magnetic sensor in the Dewar is located at each measurement position), and the magnetic field data is not immersed in each magnetic sensor. Attack is detected and collected to obtain a plurality of measurement data. [0005] Based on a plurality of obtained measurement data, the position, direction, and size of a biological activity current source related to a lesion are estimated, and the estimated current source is used as an X-ray CT apparatus or a magnetic resonance tomography apparatus. (MRI apparatus) is used to specify a physical position of an affected part or the like by displaying it on a tomographic image obtained by an (MRI apparatus) or the like. [0006] Conventionally, the current source is estimated by a least square method, a minimum norm method, a grid point moving minimum norm method, or the like. In the least square method, a current source is virtually set,
The magnetic sensor at each measurement position obtains virtual magnetic field data by calculation assuming that the small magnetic field generated by the virtual current source is measured, and the data obtained by this calculation and the data actually measured are calculated. Find the square error,
This is a method in which the virtual current source is moved so as to minimize the square error and approach a true current source. In the minimum norm method, a grid point is set in a region to be inspected, and a matrix indicating a relationship between a current source on the grid point and each measurement data is obtained.
This is a method of obtaining the current source (distribution on a grid point) by taking the product of the measurement data (for example, WHKullmann, KDJandt,
K. Rehm, HASchlitt, WJDallas and WESmith, Adva
nces in Biomagnetism pp.571-574, Plenum Pless, New
York, 1989, and the following Japanese Patent Application No. 5-160450, Japanese Patent Application No. 5-1604
No. 51, etc.). Further, the grid point moving minimum norm method is a method proposed by the present applicant in Japanese Patent Application Nos. 5-160450 and 5-160451, and the like. Of the distribution obtained by the above minimum norm method, each lattice point is moved near a large current source, and the distribution of the current source in the arrangement of the lattice points is obtained by the minimum norm method. And the same process is repeated until the minimum grid point interval becomes equal to or less than a preset convergence value, and the current source is estimated. However, the conventional example having such a configuration has the following problems. That is, when the current source is estimated by the least square method,
The quality of the estimation result can be evaluated based on the square error. That is, it can be evaluated that the estimated current source is closer to the true current source as the square error is closer to “0”. However, when noise is included in the measurement data, the magnitude of the square error does not always match the quality of the evaluation, and there is a problem that accurate evaluation cannot be performed. In addition, the minimum norm method has a problem in that it is not possible to properly judge the quality of the estimation result. Further, even in the grid point moving minimum norm method, the current source in each grid point arrangement is estimated by the minimum norm method. Therefore, it is not possible to properly judge the quality of the estimation result in each grid point arrangement. In order to estimate the final current source, the convergence determination condition is set as to whether or not the minimum grid point interval is equal to or less than a preset convergence value. However, there is a problem that the estimation method cannot have generality. The present invention has been made in view of such circumstances, and has as its object to provide a biomagnetic measurement apparatus capable of appropriately and accurately evaluating the quality of a result estimated by each estimation method. And The present invention has the following configuration in order to achieve the above object. That is, the present invention collects a plurality of measurement data by measuring a minute magnetic field generated by a biological activity current source in a diagnosis target region of a subject from a plurality of measurement positions near the diagnosis target region with a magnetic sensor. A biomagnetic measurement apparatus for estimating the position, orientation, and size of the biological activity current source based on the plurality of measurement data, wherein (a) a part of the plurality of measurement data; Estimating means for estimating the position, direction, and size of the biological activity current source using a group (hereinafter, these measurement data are referred to as “first group measurement data”); A magnetic sensor at each measurement position when measurement data of a measurement data group including measurement data not included in the measurement data (hereinafter, these measurement data is referred to as “measurement data of the second group”). Obtains, by calculation, virtual magnetic field data when it is assumed that a micro magnetic field generated by the biological activity current source estimated by the estimation means is calculated, and the data obtained by this calculation is compared with the measurement of the second group. Evaluation value calculating means for obtaining a squared error from the data and obtaining an evaluation value for determining whether the position, direction, and size of the life activity current source estimated by the estimation means are good or bad. The operation of the present invention is as follows: The estimating means uses a part of a plurality of measurement data (a first group of measurement data) to measure a biological activity current source. Is estimated by the least squares method, the minimum norm method, etc. Then, the evaluation value calculation unit calculates the measurement data group (second data) including the measurement data not included in the measurement data of the first group. Group of The virtual magnetic field data is obtained by calculation assuming that the magnetic sensor at each measurement position when measuring the measurement data of (measurement data) measures the micro magnetic field generated by the biological activity current source estimated by the estimation means. An evaluation for determining a square error between the data obtained by this calculation and the actual measurement data of the second group to determine whether the position, direction, and size of the biological activity current source estimated by the estimation means is good or bad. The measurement data of the first group and the second group
As a method of collecting the measurement data of the group, for example, all the measurement data are divided into two, and one measurement data group is set as the measurement data of the first group, and the other measurement data group is set as the measurement data of the second group. Various methods can be selected, such as setting the measurement data included in the measurement data of each group so that a part of the measurement data overlaps each other. However, the measurement data of the second group cannot be completely included in the measurement data of the first group. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a schematic configuration of a biomagnetism measuring apparatus according to an embodiment of the present invention, and FIG. 2 is a diagram showing a schematic configuration of a multi-channel SQUID sensor used in the embodiment. In the figure, reference numeral 2 denotes a magnetically shielded room, and a bed 3 on which the subject M is supine and a magnetically shielded room 2 are disposed in proximity to, for example, the head of the subject M.
A multi-channel SQUID sensor 1 for non-invasively measuring a small magnetic field generated by a biological activity current source generated in the brain of the subject is provided. Multichannel SQUID sensor 1, as shown in FIG. 2, a plurality of magnetic sensors S ₁ to S _m are housed immersed in refrigerant to the dewar in 1a. With the arrangement position of each of the magnetic sensors S _{1 to} S _m when the Dewar 1a is arranged near the head of the subject as a measurement position, each of the magnetic sensors S _{1 to} S _m performs the biological activity in the brain of the subject M. The magnetic field data generated by the current source is measured. A plurality of measurement data measured by each of the magnetic sensors S _{1 to} S _m in the multi-channel SQUID sensor 1 is supplied to a data conversion unit 4 and converted into digital data, and then collected in a data collection unit 5. Can be
The stimulating device 6 is for giving an electrical stimulus (or a sound, a light stimulus, or the like) to the subject M. The positioning unit 7 is a device for grasping a positional relationship of the head of the subject M with respect to a three-dimensional coordinate system based on the multi-channel SQUID sensor 1. For example, a plurality of portions of the head of the subject M (for example, the middle of both eyebrows and the 3
Small coils are attached to the position, and power is supplied to these small coils from the positioning unit 7. Then, by detecting the magnetic field generated from each small coil by the multi-channel SQUID sensor 1, the positional relationship of the head of the subject M with respect to the multi-channel SQUID sensor 1 is grasped. In addition, other methods for grasping the positional relationship of the head of the subject M with respect to the SQUID sensor 1 are as follows.
A method in which a light projector is attached to a to irradiate the head of the subject M with a light beam to grasp the positional relationship between the two, or disclosed in Japanese Patent Application Laid-Open Nos. 5-237065 and 6-788925. Various techniques are used. The data analyzing unit 8 corresponds to an estimating unit and an evaluation value calculating unit. Based on the measurement data collected by the data collecting unit 5, the data analyzing unit 8 is provided in the area to be examined of the subject M (in the brain). Estimate the current source,
This is for obtaining an evaluation value of the estimation result. A tomographic image obtained by, for example, an X-ray CT apparatus or an MRI apparatus is recorded on a magneto-optical disk 9 provided in association with the data analysis unit 8, and a current source estimated by the data analysis unit 8 is These evaluation images are superimposed on these tomographic images and displayed on the color monitor 10 or printed out on the color printer 11, and the evaluation values are displayed on the operation monitor 12 or the like and notified to the operator. In addition, the tomographic image obtained by the X-ray CT apparatus or the MRI apparatus is
The configuration may be such that the data is directly transmitted to the data analysis unit 8 via the communication line 13 shown in FIG. In the apparatus shown in FIG. 1, the multi-channel S
Although the magnetic field data generated from the current source is measured using the QUID sensor 1, for example, when the magnetic field data generated from the current source is measured using the single channel SQUID sensor, as shown in FIG. Then, the positional relationship between the Dewar 1a and the subject M is displaced, and the arrangement position of the magnetic sensor S _S at each position is set as a measurement position, and the magnetic sensor S _S converts magnetic field data generated from the current source at a plurality of measurement positions. It is configured to measure and collect a plurality of measurement data in the data collection unit 5. Next, the processing executed by the data analysis unit 8 will be described with reference to the flowchart shown in FIG. The flowchart in FIG. 4 shows a procedure for estimating the current source by the least square method or the minimum norm method. First, a plurality of measurement data collected by the data collection unit 5 are grouped (step S).
1). Here, as shown in FIG. 5, all measurement data D ₁
To D _m (data measured by each magnetic sensor S _{1 to} S _m )
Is divided into two, and one measurement data group (D _{1 to} D _i : i
Is a natural number not less than 2 and not more than (m-2)) as the measurement data of the first group GR1, and the remaining measurement data group (D _{i + 1}
To D _m ) are grouped as measurement data of the second group GR2. Next, a current source is estimated using the measurement data (D _{1 to} D _i ) of the first group GR1 (step S2). This estimation method is a least square method or a minimum norm method. In these methods, a current source is estimated using all measurement data (conventional example), or a current source is estimated using a part of measurement data (this method). Embodiment) The basic processing content is the same as that of the conventional example except for the difference between the embodiments. Then, the magnetic sensors (in this embodiment, the magnetic sensors S _{i + 1 to} S _{i in} this embodiment) at the respective measurement positions when the measurement data (D _{i + 1 to} D _m ) of the second group GR2 are measured.
_m ) calculates virtual magnetic field data assuming that the micro magnetic field generated by the current source estimated in step S2 is measured (step S3). This calculation method is, for example, a calculation for obtaining virtual magnetic field data assuming that each magnetic sensor measures a small magnetic field generated by a virtually set current source in the conventional least square method. The details are omitted here because they are the same. Next, the data calculated in step S3 (T
_{i + 1 to Tm} : where the suffix is the magnetic sensor S _{i + 1} to
S _m ) and the square error between the actual measurement data (D _{i + 1 to} D _m ) of the second group GR2 are obtained, and the evaluation value f is obtained (step S4). The calculation formula is shown below. ## EQU1 ## Then, the evaluation value f obtained in step S4
(Evaluation value for the estimation result of the current source obtained in step S2) is displayed on the operation monitor 12 (step S2).
5). The smaller the evaluation value f, the better the estimation result of the current source obtained in step S2. A predetermined judgment value is determined in advance, and the evaluation value f is compared with the judgment value. If the evaluation value f is larger than the judgment value, the estimation condition in step S2 (for example, In the case of the multiplication method, the initial value of the current source to be virtually set and the method of moving the current source, and the like, and in the case of the minimum norm method, the number and arrangement of the grid points are changed.
The process may be repeated from step 2 and the process may be repeated until an evaluation value f smaller than the judgment value is obtained, so that a current source closer to the true current source may be automatically obtained. As described above, in this embodiment, the first group G composed of a part of all the measurement data (D _{1 to} D _m ).
Estimating a current source with R1 of measurement data (D ₁ ~D _i), consists of the measurement data used for the estimation (D ₁ ~D _i) other than the measurement data (D _i + 1 ~D _m) Using the measurement data of the second group GR2, the first group GR1
An evaluation value f for evaluating a result (current source) estimated using the measurement data of the first group GR1 is obtained.
Even if the estimation method using the measured data of the above is the minimum norm method, the estimation result can be properly evaluated. Further, since the evaluation value f is obtained using measurement data other than the measurement data used for estimation, the conventional least-squares method has higher objectivity than the square error obtained using all the measurement data. Therefore, the quality of the estimation can be more accurately evaluated as compared with the conventional least squares method. By the way, in step S of the above-described embodiment,
In the grouping in FIG. 1, all the measurement data were divided into two, but the method of collecting the measurement data of the first group and the measurement data of the second group in the present invention is described in FIG.
It is not limited to the method of taking. That is, the measurement data of the second group GR2 is the first group GR1.
It is sufficient that the measurement data includes a measurement data group including measurement data that is not included in the measurement data. For example, as shown in FIG. 6A, a part of the measurement data of both groups GR1 and GR2 overlaps. The measurement data of each of the groups GR1 and GR2 may be determined, or as shown in FIG. 6B, all of the measurement data of the first group GR1 is completely included in the measurement data of the second group GR2. Thus, the measurement data of each group GR1 and GR2 may be determined. However, as shown in FIG. 6C, it is not possible to determine the measurement data of each of the groups GR1 and GR2 so that all of the measurement data of the second group GR2 is completely included in the measurement data of the first group GR1. Can not. This is shown in FIGS.
In (a) and (b), the tabular value f is obtained including measurement data other than the measurement data used for estimation, and an evaluation value f reflecting measurement data other than the measurement data used for estimation can be obtained. Although the estimation result can be objectively evaluated, in FIG. 6C, measurement data other than the measurement data used for the estimation is not reflected in the evaluation value f, and the estimation result may be objectively evaluated. Because you can't. In order to more objectively evaluate the estimation result (that is, more accurately evaluate), it is preferable that the measurement data of each of the groups GR1 and GR2 is completely separated as shown in FIG. Next, a process in which the data analysis unit 8 estimates a current source by the grid point moving minimum norm method will be described with reference to a flowchart shown in FIG. First, all the measurement data are grouped (step S11). This is the same as step S1 described above, and a duplicate description will be omitted. Next, grid points are set in the diagnosis target area (step S12), and a minimum norm solution (current source estimation by the minimum norm method) is obtained (step S13). Since these processes are the same as those in the conventional example, detailed description thereof will be omitted. Next, in steps S14 and S15, the evaluation value f
Ask for. This is the same as steps S3 and S4. Then, the end judgment value set in advance is compared with the obtained evaluation value f. If the evaluation value f is smaller than the end judgment value, the processing is ended. If the evaluation value f is larger than the end judgment value, steps S17 and S18 are performed. And returns to step S13 again (step S16). In step S17, an optimal movement parameter that minimizes the evaluation value f is determined. This movement parameter is a parameter for controlling the movement of a lattice point, which is described as α, β, and γ in Japanese Patent Application No. 5-160451 or the like, and is also described in the specification of the above application. As
Conventionally, this movement parameter is determined empirically. However, in this embodiment, the optimum movement parameter can be automatically determined using the evaluation value f. Specifically, a specific numerical value is substituted for the movement parameter, the lattice point is moved, and the evaluation value f at the moved lattice point is calculated. Select the value of the parameter as the optimal movement parameter. Then, in step S18, the grid point is moved using the movement parameters, and the process returns to step S13. As described above, according to this embodiment, even when the current source is estimated by the grid point moving minimum norm method, the estimation result can be properly evaluated, and the generality can be determined for the end determination. Can be provided. Further, as described above, the optimum movement parameter can be automatically determined. The evaluation value f for determining the optimum movement parameter may be obtained as follows. [0043] That is, the evaluation value f obtained by using the measurement data of the second group GR2 and f _2, conversely to this,
Estimating the current source using the measurement data of the second group GR2, and using the measurement data of the first group GR1: to evaluate the quality of the current source estimated using the measurement data of the second group GR2 The evaluation value f of
_Set to ₁ . Then, (f ₁ + f ₂ ) is used as an evaluation value f, and
Select the optimal movement parameters. With this configuration, an evaluation value using the measurement data of each of the groups GR1 and GR2 can be added, and the objectivity is improved. As shown in FIG. 8, all the measurement data are divided into three sets (gr1, gr2, gr3).
The measurement data of the set gr1 is assigned to the measurement data of the first group GR1, and the measurement data of the combination of the second set gr2 and the third set gr3 is assigned to the measurement data of the second group GR2. the evaluation of the estimation result estimated by using the measured data value f ₂₃ (second set gr2,
(Calculated using the measurement data of the third set gr3). Next, the measurement data of the second group gr2 is transferred to the first group G.
The measurement data obtained by combining the first set gr1 and the third set gr3 with the measurement data of R1 is assigned to the measurement data of the second group GR2, and the evaluation value f obtained by the same procedure as above.
₁₃ and the measurement data of the third set gr3 as the measurement data of the first group GR1, the first set gr1 and the second set g3.
The measurement data obtained by adding r2 to the second group GR2
Obtaining an evaluation value f ₁₂ obtained by assigning the measurement data. Then, an optimal movement parameter may be selected by using (f ₂₃ + f ₁₃ + f ₁₂ ) as the evaluation value f. Further, all the measurement data are divided into n sets (n is a natural number of 4 or more), and an evaluation value f for selecting an optimal movement parameter is obtained in the same manner as in the case of dividing into three sets. You may. As is apparent from the above description, according to the present invention, the current source is estimated using the first group of measurement data composed of a part of all the measurement data, and the current source is estimated. Using the measurement data of the second group including the measurement data other than the measurement data used for the estimation, an evaluation value for evaluating a result (current source) estimated using the measurement data of the first group is obtained. Therefore, even if the estimation method using the measurement data of the first group is the minimum norm method or the grid point movement minimum norm method, the estimation result can be properly evaluated. Also, this evaluation value can be used as a convergence judgment condition of the grid point moving minimum norm method,
It is possible to give generality to the grid point moving minimum norm method. Further, since the evaluation value is obtained by including data other than the measurement data used for the estimation, the conventional least squares method has higher objectivity than the square error obtained using all the measurement data. The quality of the estimation result can be accurately evaluated as compared with the conventional least squares method.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a schematic configuration of a biomagnetism measuring device according to an embodiment of the present invention. FIG. 2 shows a multi-channel SQUID used in the embodiment.
FIG. 2 is a diagram illustrating a schematic configuration of a sensor. FIG. 3 is a diagram for explaining a configuration in a case where a single-channel SQUID sensor measures a magnetic field from a current source. FIG. 4 shows a data analysis unit according to an embodiment,
9 is a flowchart illustrating a processing procedure when a current source is estimated by the minimum norm method. FIG. 5 is a diagram showing grouping of measurement data. FIG. 6 is a diagram showing a modified example of grouping of measurement data and the like. FIG. 7 is a flowchart illustrating a processing procedure in a case where the data analysis unit according to the embodiment estimates a current source by a grid point moving minimum norm method. FIG. 8 is a diagram for explaining a modification of calculation of an evaluation value. [Description of Signs] 1 Multi-channel SQUID sensor 8 Data analysis unit M Subject S _{1 to} S _m Magnetic sensors D _{1 to} D _m Measurement data GR 1 First group GR 2 Second group

Claims (1)

(57) [Claims 1] A minute magnetic field generated by a biological activity current source in a diagnosis target region of a subject is measured by a magnetic sensor from a plurality of measurement positions near the diagnosis target region. Collect multiple measurement data, and based on the multiple measurement data,
In the biomagnetic measurement apparatus for estimating the position, direction, and size of the biological activity current source, (a) a part of a plurality of measurement data groups (hereinafter, these measurement data are referred to as “ Measurement data of the first group ”)
Estimating means for estimating the position, orientation, and size of the biological activity current source by using: (b) a measurement data group including measurement data not included in the measurement data of the first group (hereinafter, these measurement data When it is assumed that the magnetic sensor at each measurement position when measuring the measurement data of “the second group of measurement data” measures the minute magnetic field generated by the biological activity current source estimated by the estimation means. Virtual magnetic field data is obtained by calculation, the square error between the data obtained by this calculation and the measurement data of the second group is obtained, and the position, direction, and magnitude of the biological activity current source estimated by the estimation means are obtained. A biomagnetism measuring device comprising: an evaluation value calculating means for obtaining an evaluation value for judging the quality of the biomagnetism.