JP7306070B2 - SIGNAL PROCESSING DEVICE, SIGNAL PROCESSING METHOD AND SIGNAL PROCESSING PROGRAM - Google Patents

Description

The present invention relates to a signal processing device, a signal processing method, and a signal processing program.

For example, as a method for measuring a biomagnetic field, a method is known in which a part of a subject is stimulated to induce neural activity in a region to be measured, and a magnetic field emitted from this neural activity is measured by a sensor. In this type of measurement method, an interfering magnetic field induced by the stimulus may be generated by the stimulus or the movement of the muscle caused by the stimulus, resulting in noise.

Therefore, there is a method of acquiring a signal of interest that does not include interfering magnetic field data by removing components of measurement data that includes interfering magnetic field data and does not include magnetic field data of interest from the components of measurement data that includes interfering magnetic field data. Proposed. In this method, first, measurement data including interfering magnetic field data is obtained by applying stimulation while the measurement target region is brought close to the magnetic field measurement device, and magnetic field data of the target of interest is obtained while the measurement target region is separated from the magnetic field measurement device. Measurement data that does not include Then, the signal of interest is acquired by removing the component of the next acquired measurement data from the component of the first acquired measurement data (Patent Document 1).

In addition, a body motion signal filtered by a transfer function calculated by modeling the effect of body motion on blood flow is input to an adaptive filter, and the output signal of the adaptive filter is subtracted from the pulse wave signal to obtain a pulse wave signal. A technique for removing body motion noise from the image has been proposed (Patent Document 2).

When obtaining magnetic field data of interest by removing interfering magnetic field data based on two pieces of measurement data, it is necessary to perform measurement twice, which poses a problem of prolonging examination time. In addition, if the signal source of the interfering magnetic field is near or within the measurement area of the magnetic field data, there is a risk that the magnetic field component to be measured and the interfering magnetic field component contained in the measurement data cannot be distinguished. There is Thus, if the interfering magnetic field data cannot be removed from the measurement data, no valid magnetic field data of interest can be obtained.

An object of the present invention is to reduce the measurement time while removing interfering magnetic field data from the magnetic field measurement data and extracting the magnetic field data of interest.

In order to solve the above technical problem, a signal processing apparatus according to one embodiment of the present invention comprises a signal of interest generated by a nerve-induced magnetic field generated by nerve activity of a subject induced by electrical stimulation , and a measurement execution unit that acquires measurement data including an interfering signal generated by an interfering magnetic field generated in the vicinity of a signal source of the signal of interest; and a signal source of the signal of interest and a signal of the interfering signal based on the measurement data. and an interfering signal source for extracting interfering signal data generated from the signal source of the interfering signal based on the result of estimating the signal source by the signal source estimating unit. an extraction unit; and a target signal extraction unit for extracting a signal of interest by removing a common portion between the measurement data acquired by the measurement execution unit and the interfering signal data , wherein the signal source of the interfering signal is , a site that is present in the subject other than the measurement target site that generates the signal of interest and that generates an interfering magnetic field based on electrical stimulation for inducing nerve activity, or the electrical stimulation electrodes attached to the subject for providing the subject with a

It is possible to reduce the measurement time while extracting the magnetic field data of interest by removing the interfering magnetic field data from the measured magnetic field data.

1 is a diagram showing a main part of a biomagnetic field measuring device including a signal processing device according to a first embodiment; FIG. FIG. 2 is a diagram showing a scene of measurement of a magnetic field generated from a subject by a biomagnetic field measuring device; 1 is a block diagram showing an example of a signal processing device according to a first embodiment; FIG. 4 is a diagram illustrating an example of a hardware configuration of the signal processing device of FIG. 3; FIG. 4 is a flow diagram showing an example of the operation of the signal processing device of FIG. 3; FIG. FIG. 4 is a diagram showing an example of measurement data acquired by measurement by the measurement execution unit in FIG. 3; FIG. FIG. 4 is a diagram showing an example of interference signal data (magnetic field components) generated by the interference signal source extraction unit of FIG. 3; FIG. 4 is a diagram showing an example of a target signal of interest extracted by the target signal extraction unit of FIG. 3; FIG. 9 is an enlarged view of the main part of FIG. 8; FIG. 10 is a block diagram showing an example of a signal processing device according to a second embodiment; FIG. 11 is a flow chart showing an example of the operation of the signal processing device of FIG. 10; FIG. FIG. 11 is a block diagram showing an example of a signal processing device according to a third embodiment; FIG. FIG. 11 is a block diagram showing an example of a signal processing device in a fourth embodiment; FIG. FIG. 14 is a diagram showing an example of measurement data acquired by the measurement performed by the measurement execution unit shown in FIG. 3, FIG. 10, FIG. 12, or FIG. 13; FIG. 14 is a diagram showing an example of a waveform of a signal of interest extracted by the signal-of-interest extraction unit shown in FIG. 3, 10, 12, or 13 using a method based on the SSP method; 16 is an enlarged view of the inside of the dashed frame in FIG. 15; FIG. FIG. 10 is a diagram showing examples of magnetic field data measured in the palm of a hand without applying artifact removal, applying the present invention, and applying a conventional method. FIG. 18 is a diagram showing an example of visualization of nerve activity currents for the magnetic field data shown in FIG. 17; FIG. 4 is a diagram showing an example of setting current waveform acquisition points at equal intervals on the conduction path of an axonal current component; FIG. 20 is a diagram showing examples of current waveforms acquired at respective current waveform acquisition points in FIG. 19; FIG. 4 is a diagram showing a magnetic field waveform of a sensor output that particularly clearly captures each biomagnetic field signal in the magnetic field data measured by the palm.

Embodiments will be described below with reference to the drawings.
(First embodiment)
FIG. 1 is a diagram showing a main part of a biomagnetic field measuring device 100 including a signal processing device 10 (FIG. 2) according to the first embodiment.

The biomagnetic field measurement device 100 has a base portion 102 including a contact portion with the subject P, a magnetic field detection portion 104 arranged in the base portion 102, and a nerve stimulation device (not shown) connected to the electrodes 106. ing. The magnetic field detection unit 104 is an example of a measurement execution unit. The nerve stimulator applies electrical stimulation to the subject P via the electrodes 106 according to predetermined stimulation conditions. In FIG. 1 the electrodes 106 are obscured by a band for attachment to the subject P. FIG.

The magnetic field detection unit 104 includes a sensor array including a plurality of sensors that detect magnetic fields and a drive circuit that drives the sensor array. FIG. 1 shows only the sensor array. The magnetic field detection unit 104 measures the magnetic field generated by the electrical stimulation applied to the subject P, which is the magnetic field created by the electrical signals in the body that indicate the movement of nerves in the measurement target region of the subject P.

In the biomagnetic field measurement apparatus 100, the subject P (such as the palm) is brought into contact with the measurement area 108 where the magnetic field can be measured by the magnetic field detection unit 104, and the subject P is stimulated from the electrodes. The sensor array measures the magnetic field generated by the nerve activity of the measurement target site. In FIG. 1, the palm of the subject P is the measurement target site, and the fingertip (near the first joint of the middle finger) is brought into contact with the stimulating electrode to apply electrical stimulation. A sensor array acquires a nerve-induced magnetic field generated by nerve activity.

In this case, the electrode 106 that is in contact with the middle finger and an electric cable (not shown) connected to the electrode are the source of the interfering signal (interfering magnetic field). For example, an interfering signal is noise (stimulation artifact) generated outside the subject P to be measured in response to a stimulus. However, since the electric cable is generally farther from the magnetic sensor than the electrode 106 , the interfering signal obtained by the magnetic field detector 104 is predominantly generated from the electrode 106 . For this reason, the description below assumes that the interfering signal is generated from the electrode 106 .

It should be noted that the stimulus given to the subject P may be something other than current, and for example, may be given to the measurement target site by magnetism. Alternatively, electrodes may be brought into contact with both the middle finger and the index finger, and electrical stimulation may be applied to both.

FIG. 2 is a diagram showing a scene of measurement of the magnetic field generated from the subject P by the biomagnetic field measurement device 100. As shown in FIG. The palm of the subject P is placed on the pedestal 102 in a measurement area 108 facing the sensor array arranged in the pedestal 102 . In the example shown in FIG. 2 , the electrode 106 that stimulates the subject P and is the signal source of the interfering signal is located outside the measurement area 108 .

The extraction target region 110 is a region including the position of the palm, which is the signal source of the signal of interest, and the position of the electrode 106, which is the signal source of the interfering signal. set. As will be described later, the signal processing device 10 (FIG. 3) of the present embodiment estimates the current distribution in the extraction target region 110, and detects the magnetic field generated from the signal source of the interference signal and detected by the sensor array. The data (interfering signal data) are extracted by computation. Then, the signal processing device 10 removes the common portion between the measurement data and the interference signal data acquired by the magnetic field detection unit 104, thereby extracting the target signal of interest emitted from the palm, which is the part to be measured.

In FIG. 2, the palm of the subject is the target of measurement, but any part of the subject may be the target of measurement. Also, the subject is not limited to humans, and may be an animal such as an ape.

FIG. 3 is a block diagram showing an example of the signal processing device 10 according to the first embodiment. Note that FIG. 3 also shows functional blocks of the signal processing device 10 . As described above, the signal processing device 10 is included in the biomagnetic field measurement device 100 and processes magnetic field data acquired by the biomagnetic field measurement device 100 . The signal processing device 10 has a measurement executing section 11 , a signal source estimating section 12 , an interfering signal source extracting section 13 and a target signal extracting section 14 .

For example, the measurement execution unit 11 includes the magnetic field detection unit 104 (sensor array) shown in FIG. is realized by the control function of The control function of the magnetic field detection unit 104 by the measurement execution unit 11 may be realized by a logic circuit such as an FPGA (Field Programmable Gate Array).

The measurement execution unit 11 acquires measurement data including magnetic field information emitted from the measurement target site of the subject P in a state in which the measurement target site is in close contact with the sensor array. The measurement data includes a signal of interest indicating a magnetic field to be measured emitted from a site to be measured, and an interfering signal generated in the vicinity of the sensor array, which is the signal source of the signal of interest. The measurement data acquired by the measurement execution unit 11 may be temporarily stored in the memory within the signal processing device 10 in order to be used for subsequent processing by the target signal extraction unit 14 .

For example, the interfering signal is generated by the magnetic field generated by the electrical stimulation given to the subject P, and is generated from the electrode 106 attached to the subject P to apply the electrical stimulation and the body tissue in the vicinity thereof. Therefore, in the following description, it is assumed that the signal source of the interfering signal is the electrode 106 .

For example, the measurement execution unit 11 can acquire measurement data _BS according to the model shown in Equation (1).
B _S =A+B+ε (1)
In equation (1), symbol A denotes the interfering signal component, symbol B the signal component of interest, and symbol ε the white noise.

For example, the functions of the signal source estimator 12, the interfering signal source extractor 13, and the target signal extractor 14 are implemented by a signal processing program executed by a controller such as a CPU included in the signal processor 10. FIG. The functions of the signal source estimation unit 12, the interfering signal source extraction unit 13, and the target signal extraction unit 14 may be realized by logic circuits such as FPGA.

The signal source estimator 12 estimates a signal source that generates a magnetic field in an extraction target region 110 that includes the sensor array that is the signal source of the signal of interest and the electrode 106 that is the signal source of the interfering signal. When estimating the signal source, first, an extraction target region 110 is set that includes both the position of the signal source of the signal of interest and the position of the signal source of the interfering signal. The signal source estimator 12 estimates the current distribution in the plane of the extraction target region 110 using an estimation algorithm such as the spatial filter method (Non-Patent Document 1).

Note that when the signal source of the interfering signal is distant from the signal source of the signal of interest and it is difficult to set the extraction target region 110 including the signal source of the interfering signal due to the positional relationship with the sensor array, the signal of interest An extraction target region 110 that includes only the position of the source may be set. However, in this case, it is preferable to expand the extraction target region 110 toward the position of the signal source of the interfering signal. Further, for example, using the method of Patent Document 3, the extraction target region 110 is set so as to include the position of the signal source of the interfering signal based on the sensor array and the morphological information indicating the position of the signal source of the interfering signal. may

For example, the interfering signal source extraction unit 13 selects an area having interfering signal characteristics from among the signal sources estimated by the signal source estimating unit 12 as the signal source of the interfering signal. Morphological information indicating the position of the signal source of the interfering signal may be used to detect the position of the signal source of the interfering signal.

The selection of the signal source of the interfering signal may be automatically performed using the previously recognized positional information of the signal source of the interfering signal. Alternatively, the selection of the signal source of the interfering signal may be performed by displaying a current map or the like that allows the current distribution to be visually recognized on the display device and having the operator of the biomagnetic field measurement device 100 input the position of the signal source of the interfering signal. good.

Then, the interfering signal source extraction unit 13 extracts interfering signal data (current component) generated from the selected signal source of the interfering signal. For example, the jamming signal data is extracted by determining the temporal transition of the current of the signal source of the jamming signal based on the current distribution. Here, in biomagnetic field measurement, it is considered that the relationship between the change in the current component of the virtual interference signal and the change in the magnetic field component generated from the selected signal source is linear. Therefore, the interference signal source extraction unit 13 generates virtual interference signal data (magnetic field component) by, for example, multiplying the extracted interference signal data (current component) by a predetermined coefficient. The virtual interfering signal data is output to the target signal extraction unit 14 . Here, the virtual jamming signal data (magnetic field component) is the predicted value of the magnetic field data generated by the signal source of the jamming signal.

The virtual interfering signal data (magnetic field component) _Ba generated by the interfering signal source extractor 13 is considered to follow equation (2).
B _a =A+C+ε (2)
In equation (2), A denotes an interfering signal component, C denotes a signal component having a distribution different from that of the signal of interest, and ε denotes white noise. Equation (2) is the same as Equation (1) except that instead of not having the signal of interest component B of Equation (1), it has a signal component C with a different distribution than the signal of interest.

The target signal extraction unit 14 extracts the target signal by removing the common portion between the measurement data acquired by the measurement execution unit 11 and the virtual interference signal data. "+" shown above and "-" shown below in the target signal extraction unit 14 in FIG. 3 indicate the concept of extracting the target signal by subtracting the component of the virtual interference signal data from the component of the measurement data. there is

For example, the technique of Non-Patent Document 3 is used to remove the common portion between the measurement data and the virtual interference signal data. As a result, it is possible to extract the signal component of interest from which the interfering signal component has been removed from the measurement data by performing one-time measurement of the part to be measured on the subject P. FIG.

FIG. 4 is a diagram showing an example of the hardware configuration of the signal processing device 10 of FIG. 3. As shown in FIG. The signal processing device 10 is an information processing device, for example, and has a CPU 21 , a RAM 22 , a ROM 23 , an auxiliary storage device 24 , an input/output interface 25 and a display device 26 , which are interconnected by a bus 27 .

The CPU 21 controls the overall operation of the signal processing device 10 . The CPU 21 implements various functions shown in FIG. 3 by executing a signal processing program stored in the ROM 23 or the auxiliary storage device 24 . Note that the CPU 21 may control the operation of the biomagnetic field measurement device 100 as a whole.

The RAM 22 is used as a work area for the CPU 21 and may include a non-volatile RAM that stores signal processing programs and information. The ROM 23 stores various programs and parameters used by the various programs. A signal processing program of the present invention may be stored in the ROM 23 .

The auxiliary storage device 24 is a storage device such as an SSD (Solid State Drive) or HDD (Hard Disk Drive). Various data, files, etc. required for the operation of the device 10 are stored.

The input/output interface 25 includes user interfaces such as a touch panel, keyboard, operation buttons, and speakers, and a communication interface for communicating with other electronic devices. The display device 26 displays an operation window for causing the measurement execution unit 11 to perform measurement, a waveform representing the measurement data acquired by the measurement execution unit 11, and the like.

FIG. 5 is a flow chart showing an example of the operation of the signal processing device 10 of FIG. That is, FIG. 5 shows an example of a signal processing method by the signal processing device 10 and a signal processing program for causing the signal processing device 10 to execute signal processing.

First, in step S10, the measurement execution unit 11 acquires measurement data including magnetic field information emitted from the site to be measured. Next, in step S20, the signal source estimator 12 estimates a signal source that generates a magnetic field in the extraction target region 110 shown in FIG. 2 and obtains a current distribution.

Next, in step S30, the interfering signal source extraction unit 13 extracts, as interfering signal data (current components), signal sources in a region considered to be the source of the interfering signal from the signal sources estimated in step S20. . The interference signal source extraction unit 13 generates virtual interference signal data (magnetic field component) based on the extracted interference signal data (current component).

Next, in step S40, the target signal extraction unit 14 extracts the target signal by removing the common portion between the measurement data acquired by the measurement execution unit 11 and the virtual interference signal data. The process of removing the virtual interfering signal data from the measured data to generate the signal of interest is then completed.

FIG. 6 is a diagram showing an example of measurement data obtained by measurement by the measurement execution unit 11 of FIG. FIG. 6 shows an example of time change (raw data) of the magnetic flux density, which is the output value of each sensor in the sensor array, and electrical stimulation is applied 5 ms after the start of measurement. FIG. 6 shows measurement data after electrical stimulation was applied.

As shown in FIG. 2, the sensor array (measurement area 108) and the electrode 106, which is the stimulation application site, are separated by only about several centimeters. In FIG. 6, from the position of the hand and the conduction velocity of the induced nerve activity, it is considered that the magnetic field waveform emitted by the current due to nerve activity exists in the vicinity of 6 ms to 10 ms. However, the magnetic field waveform associated with nerve activity is buried in artificial noise (artifact noise) generated by electrical stimulation, and cannot be confirmed in FIG.

FIG. 7 is a diagram showing an example of interfering signal data (magnetic field components) generated by the interfering signal source extractor 13 of FIG. The scale of FIG. 7 is the same as that of FIG. The interference signal data shown in FIG. 7 is included in the measurement data shown in FIG. 6, and it can be seen that the artifact noise generated from the electrodes 106 due to electrical stimulation is considerably large.

8 and 9 are diagrams showing examples of target signals of interest extracted by the target signal extraction unit 14 of FIG. 8 and 9 are generated by removing the common portion between the measurement data (FIG. 6) acquired by the magnetic field detection unit 104 and the interference signal data (FIG. 7) generated by the interference signal source extraction unit 13. 4 shows the change in the magnetic field of the signal of interest. 8 is the same scale as FIGS. 6 and 7. FIG. FIG. 9 is an enlarged view of the main part of FIG. 8, and the scale of the magnetic flux density (vertical axis) is 1000 times that of FIG.

As can be seen from FIGS. 8 and 9, the processing by the signal processing device 10 extracts the signal of interest buried in the measurement data due to the interfering signal data using the measurement data acquired in one measurement. can be done. That is, by predicting the jamming signal data (magnetic field component) detected by the sensor array based on the prediction of the current component at the signal source of the jamming signal data without performing measurement for the purpose of acquiring jamming signal data, A signal of interest can be extracted.

As described above, in this embodiment, the signal of interest can be extracted from the measurement data including the interfering signal data without performing the measurement for the purpose of acquiring the interfering signal data. As a result, since the number of measurements can be reduced to one, the measurement time (that is, inspection time) can be shortened compared to the conventional technique, and the burden on the subject P can be reduced.

Also, in this embodiment, the subject P is required to maintain the same pose for, for example, several minutes during the measurement in order to extract a good quality signal of interest. If the subject P moves during measurement, the measurement data may contain noise components associated with the movement of the subject P and cannot be used. Shortening the measurement time by extracting the signal of interest in a single measurement leads to shortening the time required to maintain the same posture of the subject P, and as a result, it is possible to stably measure a high-quality signal of interest. leads to

(Second embodiment)
FIG. 10 is a block diagram showing an example of a signal processing device 10B according to the second embodiment. Elements that are the same as those in FIG. 3 are given the same reference numerals, and detailed description thereof is omitted. FIG. 10 also shows functional blocks of the signal processing device 10B. The signal processing device 10B has an interfering signal source extraction unit 13A instead of the interfering signal source extraction unit 13 in FIG. there is

The interference signal source extractor 13A extracts interference signal data (current component) generated from the selected signal source of the interference signal, and outputs the extracted interference signal data to the virtual interference signal generator 15A. The interference signal source extraction unit 13A has the same function as the interference signal source extraction unit 13 in FIG. 3 except that it does not have the interference signal data (magnetic field component) generation function of the interference signal source extraction unit 13 in FIG. .

The virtual interfering signal generator 15A calculates virtual interfering signal data (magnetic field component) expected to be obtained from the sensor array based on the interfering signal data (current component) extracted by the interfering signal source extractor 13A. That is, the virtual interfering signal generation unit 15A generates predicted values of the interfering signal data measured by the sensor array, which is the magnetic field data generated by the signal source of the interfering signal, based on the current distribution at the signal source of the interfering signal. calculate. The process of extracting the target signal of interest based on the measurement data and the virtual interfering signal data by the target signal extraction unit 14 is the same as in the first embodiment.

FIG. 11 is a flow chart showing an example of the operation of the signal processing device 10A of FIG. That is, FIG. 11 shows an example of a signal processing method by the signal processing device 10A and a signal processing program that causes the signal processing device 10A to execute signal processing. A detailed description of the same processing as in FIG. 5 will be omitted.

In the process shown in FIG. 11, step S30A is performed instead of step S30 of FIG. 5, and step S35A is inserted between step S30A and step S40. The processing of steps S10, S20, and S40 is the same as in FIG.

In step S30A, the interfering signal source extractor 13A extracts, as interfering signal data (current component), a signal source in a region considered to be an interfering signal source from among the signal sources estimated in step S20. Then, the interference signal source extractor 13A outputs the extracted interference signal data (current component) to the virtual interference signal generator 15A.

Next, in step S35A, the virtual interfering signal generator 15A calculates virtual interfering signal data (magnetic field component) expected to be obtained by the sensor array based on the interfering signal data (current component), and extracts the target signal. Output to the unit 14 . Thereafter, the same processing as in step S40 of FIG. 5 is performed, and the signal of interest is extracted by removing the common portion between the measurement data and the virtual interference signal data, and the processing ends.

As described above, the same effects as in the first embodiment can be obtained in the second embodiment. Furthermore, in the second embodiment, by outputting the virtual interference signal data (magnetic field component) calculated by the virtual interference signal generation unit 15A to the target signal extraction unit 14, the target signal extraction unit 14 can generate a highly accurate virtual interference signal. Signal data can be used to extract a signal of interest. As a result, the accuracy of inspection can be further improved.

(Third embodiment)
FIG. 12 is a block diagram showing an example of a signal processing device 10B according to the third embodiment. Elements that are the same as those in FIG. 3 are given the same reference numerals, and detailed description thereof is omitted. FIG. 12 also shows functional blocks of the signal processing device 10B. The signal processing device 10B is the same as the signal processing device 10 of FIG. 3 except that it has an interference signal source extraction section 13B instead of the interference signal source extraction section 13 of FIG.

Interfering signal source extraction unit 13B receives information indicating the mutual positional relationship of various signal sources in extraction target region 110 shown in FIG. A region having characteristics of the interfering signal is selected from among the signal sources as the signal source of the interfering signal. The information received by the interfering signal source extraction unit 13B includes the positional information of the palm of the subject P, which is the signal source of the signal of interest, and the positional information of the electrode 106, which is the signal source of the artifact noise.

Then, the interfering signal source extraction unit 13B extracts interfering signal data (current component) generated from the selected interfering signal source, and extracts virtual interfering signal data, which is a predicted value of the magnetic field data generated by the interfering signal source. (magnetic field component). The generated virtual interfering signal data is output to the target signal extraction section 14 . By using information indicating the mutual positional relationship of various signal sources, it is possible to improve the accuracy of the virtual interfering signal data (current component and magnetic field component) generated by the interfering signal source extractor 13B.

As described above, the same effect as in the first embodiment can be obtained in the third embodiment. Furthermore, in the third embodiment, the accuracy of interference signal data (current components and magnetic field components) can be improved by using information indicating mutual positional relationships of various signal sources. As a result, the accuracy of extraction of the target signal of interest by the target signal extraction unit 14 can be increased, and the accuracy of inspection can be further improved.

(Fourth embodiment)
FIG. 13 is a block diagram showing an example of a signal processing device 10C according to the fourth embodiment. Elements that are the same as those in FIG. 10 are given the same reference numerals, and detailed description thereof will be omitted. FIG. 13 also shows functional blocks of the signal processing device 10C. The signal processing device 10C is the same as the signal processing device 10A of FIG. 10 except that it has an interference signal source extraction section 13C instead of the interference signal source extraction section 13A of FIG.

Interference signal source extraction unit 13C receives information indicating the mutual positional relationship of various signal sources in extraction target region 110 shown in FIG. A region having characteristics of the interfering signal is selected from among the signal sources as the signal source of the interfering signal. Then, the interference signal source extractor 13C extracts interference signal data (current component) generated from the selected signal source of the interference signal, and outputs the extracted interference signal data to the virtual interference signal generator 15A.

The interfering signal source extraction unit 13C can correctly select the signal source of the interfering signal by using the information indicating the mutual positional relationship of the various signal sources, and accurately extract the interfering signal data (current component). be able to. As a result, the accuracy of the virtual interfering signal data generated by the virtual interfering signal generator 15A can be improved, the extraction accuracy of the signal of interest can be increased, and the inspection accuracy can be further improved.

As described above, the same effects as those of the second embodiment can be obtained in the fourth embodiment. Furthermore, in the fourth embodiment, by using information indicating the mutual positional relationship of various signal sources, interference signal data (current components) can be extracted with high accuracy, and the extraction accuracy of the signal of interest can be improved. Therefore, the accuracy of inspection can be further improved.

In any of the above-described embodiments, the target signal extraction unit 14 uses the interference signal data or the virtual interference signal data for the input measurement data, the SSP method (Signal Subspace projection method; Non-Patent Document 4) ) may be used to extract the signal of interest. The SSP method is a technique for removing components contained in jamming signal data or virtual jamming signal data, and can correctly extract the signal of interest if the jamming signal data model follows Equation (3) below.
B _i =A+ε (3)
In equation (3), symbol A denotes an interfering signal component, symbol ε denotes white noise, and virtual interfering signal data B _i is obtained by adding only interfering signal component _A to virtual interfering signal data Ba obtained by equation (2). This is the data when the is included.

Figures 14-16 show example data for extracting a signal of interest using the SSP method.

FIG. 14 is a diagram showing an example of measurement data acquired by measurement by the measurement execution unit 11 shown in FIG. 3, FIG. 10, FIG. 12, or FIG. In FIG. 14, similarly to FIG. 6, the interfering signal generated by the electrical stimulation can be mainly confirmed, but the biological signal component (ie, the signal of interest) cannot be confirmed because it is buried in the interfering signal.

FIG. 15 is a diagram showing an example of a waveform of a signal of interest extracted by the signal-of-interest extractor 14 shown in FIG. 3, 10, 12 or 13 using the SSP method. FIG. 16 is an enlarged view of the inside of the broken line frame in FIG. 15 . In FIGS. 15 and 16, as can be seen by comparing with FIG. 14, the biosignal, which is the signal of interest, can be confirmed with the interference signal component removed.

In the above-described embodiment, as shown in FIG. 2, the case where the electrode 106, which is the signal source of the interfering signal, is outside the measurement area 108 has been described. However, the source of the jamming signal may be within the measurement region 108, and the signal of interest can still be extracted by removing the intersection of the measured data and the virtual jammer data.

This is because the signal source estimating unit 12 estimates the position of the electrode 106 or the like, which is the signal source of the interfering signal, based on the current distribution in the extraction target region 110. Therefore, if the current distribution is obtained, the signal source of the interfering signal is This is because the position of the signal source can be estimated even when it is within the measurement area 108 . Therefore, the interference signal source extraction unit 13, 13A, 13B or 13C can select the signal source of the interference signal data and extract the interference signal data, and the target signal extraction unit 14 can extract the signal of interest. can.

Also, in the above-described embodiments, a method of extracting a signal of interest by removing interfering signal data due to artifact noise from measurement data has been described. However, the interfering signal data to be removed from the measurement data may be biomagnetic field data (biomagnetic field noise) other than the biomagnetic field data generated at the measurement target site in response to stimulation. For example, magnetic field data generated by muscle activity can be cited as biomagnetic field data generated outside the measurement target site. In the above-described embodiment, the position of the electrode 106 or the like, which is the signal source of the interfering signal, is estimated based on the current distribution in the extraction target region 110. Therefore, for example, in one measurement, both the artifact noise and the biomagnetic noise can be removed.

(Application example)
A specific example to which the technique of the present invention described above is applied will be described below. In addition, although an example in which biomagnetic field measurement is performed using the palm as a measurement target site is described, the present invention is not limited to this, and the method of the present invention can be applied to other sites.

As shown in FIG. 2, the subject's palm is placed on the device, electrical stimulation is applied from the stimulating electrode 106, the magnetic field generated by the induced nerve activity current is measured multiple times, and white noise is reduced by averaging. magnetic field data was acquired. Magnetic field data obtained without artifact removal, magnetic field data to which the present invention has been applied, and magnetic field data to which the conventional method has been applied are shown. The magnetic field data are shown in FIG. 17, with the output waveforms from all the sensors plotted along the horizontal axis as time and the vertical axis as magnetic field intensity.

In the magnetic field data to which the artifact removal method was not applied, the stimulus artifact remains, and the biomagnetic field signal cannot be confirmed. On the other hand, in the magnetic field data to which the present invention is applied, the biomagnetic signal can be confirmed before the latency of 5 ms. In the magnetic field data using the conventional method, the biomagnetic signal can be confirmed before the latency of 5 ms. I can confirm.

The RENS filter ( Non-Patent Document 5) was applied to visualize nerve activity currents. The visualized nerve activity current is superimposed on the morphological information using the technique of Patent Document 3, and is shown in FIG.

In FIG. 18, the intensity of the current is shown as the brightness of the contour map, and the closer the contour line is to white, the stronger the current is. Furthermore, the direction of the current at each position within the measurement object is shown as a gray arrow.

Nerve activity current compensates for two intra-axonal current components along the axon: a leading axonal current component pointing in the conduction direction from the depolarization and a trailing axonal current component pointing in the opposite conduction direction from the depolarization. It consists of volume current components flowing outside the nerve axon, and each current component conducts while maintaining the positional relationship.

Confirming the current distribution visualized from the data without artifact removal, only strong currents exist on the distal and lateral sides of the hand, and no neural activity current component is observed.

When confirming the current distribution visualized from the data after the application of the present invention, it can be seen that the nerve activity current propagates proximally from the middle finger to which the stimulation was applied.

When confirming the current distribution visualized from the data after applying the conventional method, it can be seen that the neural activity current propagates proximally from the middle finger to which the stimulation was applied. However, d. When confirming the current distribution visualized from the magnetic field data with a latency of 3.6 ms, current components other than the nerve activity current are observed on the distal side of the hand. This is considered to be the artifact component that could not be completely removed visualized as a current.

In addition, since the current distribution obtained by the spatial filtering method has position information and current intensity for each time point, it is possible to acquire the current waveform at any point.

A plurality of current waveform acquisition points are set at equal intervals on the conduction path of the axonal current component as shown in FIG. In the current waveform obtained from the data to which the present invention is applied and the current waveform obtained from the data to which the conventional method is applied, the closer the current waveform acquisition point is, the more the peak latency is shifted backward on the time axis. , it has been shown that the conduction of nerve activity can also be acquired as a waveform. On the other hand, in the current waveform obtained from the data to which the artifact removal was not applied, the nerve activity current waveform was buried in the artifact current waveform, and the conduction of nerve activity could not be evaluated as a waveform.

By applying the present invention, it was confirmed that magnetic field data with artifacts removed can be obtained in half the time of the conventional method, and neural activity can be evaluated as a waveform comparable to the case of using the conventional method.

In [0040] and [0050], a case of using a magnetic field component and an embodiment using a current component as interfering signal data to the target signal extractor were described, respectively. FIG. 21 shows the magnetic field data obtained by applying the methods described in [0040] and [0050] to the magnetic field data obtained by the method described in [0073]. The output waveforms of the five sensors (whose waveform shapes could be clearly confirmed) are shown side by side. The waveform obtained by applying the method described in [0040] is shown as a solid black line, and the waveform obtained by applying the method described in [0050] is shown as a black dashed line. It can be confirmed that comparable effects are obtained by using either method. Note that the unit of the horizontal axis of each graph is [ms].

Although the present invention has been described above based on each embodiment, the present invention is not limited to the requirements shown in the above embodiments. These points can be changed without impairing the gist of the present invention, and can be determined appropriately according to the application form.

10, 10A, 10B, 10C signal processing device 11 measurement execution unit 12 signal source estimation unit 13, 13A, 13B, 13C interference signal source extraction unit 14 target signal extraction unit 15A virtual interference signal generation unit 21 CPU
22 RAMs
23 ROMs
24 Auxiliary storage device 25 Input/output interface 26 Display device 27 Bus 100 Biomagnetic field measurement device 102 Base 104 Magnetic field detection unit 106 Electrode 108 Measurement region 110 Extraction target region P Subject

JP 2018-192236 A JP 2016-221092 A WO2016/175020

K. Sekihara, S. S. Nagarajan, Adaptive Spatial Filters for Electromagnetic Brain Imaging, 2008, Springer K. Sekihara, et al., Dual signal subspace projection (DSSP): A novel algorithm for removing large interference in biomagnetic measurements, 2016, J. Neural Eng., 13, 036007 T. Watanabe, et al., Removal of Stimulus-Induced Artifacts in Functional Spinal Cord Imaging, 2013, 35th Annual International Conference of the IEEE EMBS M. A. Uusitalo, R. J. Ilmoniemi, Signal-space projection method for separating MEG or EEG into components, Medical & Biological Engineering & Computing 135-140, 1997 Sekihara K, Nagarajan S. S. Electromagnetic Brain Imaging: A Bayesian Perspective.: Springer International Publishing; 2015.

Claims (18)

a signal of interest generated by a nerve-induced magnetic field generated by nerve activity of a subject induced by electrical stimulation ; a measurement execution unit that acquires measurement data including
a signal source estimator for estimating a signal source in an extraction target region including the signal source of the signal of interest and the signal source of the interfering signal based on the measurement data;
an interfering signal source extraction unit that extracts interfering signal data generated from the signal source of the interfering signal based on the estimation result of the signal source by the signal source estimating unit;
a target signal extraction unit that extracts a target signal of interest by removing a common portion between the measurement data acquired by the measurement execution unit and the interfering signal data ;
The signal source of the interfering signal exists at a site other than the measurement target site that generates the signal of interest in the subject, and generates an interfering magnetic field based on electrical stimulation for inducing neural activity. site or electrode attached to the subject for providing the electrical stimulation to the subject
A signal processing device characterized by : a measurement execution unit that acquires measurement data including a signal of interest generated from a subject and an interfering signal generated in the vicinity of the signal source of the signal of interest;
a signal source estimator for estimating a signal source in an extraction target region including the signal source of the signal of interest and the signal source of the interfering signal based on the measurement data;
an interfering signal source extraction unit that extracts interfering signal data generated from the signal source of the interfering signal based on the estimation result of the signal source by the signal source estimating unit;
a target signal extraction unit that extracts a target signal of interest by removing a common portion between the measurement data acquired by the measurement execution unit and the interfering signal data ;
The signal processing apparatus, wherein the target signal extraction unit extracts the target signal using a signal subspace projection method . The measurement data is magnetic field data,
The signal source estimating unit obtains a current distribution in the extraction target region, estimates a signal source of the interference signal based on the obtained current distribution,
The interference signal source extraction unit extracts a magnetic field component of the interference signal data generated from the signal source of the interference signal estimated based on the current distribution.
3. The signal processing apparatus according to claim 1 or 2 , characterized by : a virtual interfering signal generating unit that generates virtual interfering signal data from the interfering signal data extracted by the interfering signal source extracting unit;
The target signal extracting unit uses the virtual interfering signal data instead of the interfering signal data, and removes a common portion between the measurement data acquired by the measurement executing unit and the virtual interfering signal data, thereby extracting a signal of interest. The signal processing device according to any one of claims 1 to 3, characterized in that it extracts The virtual interference signal generation unit generates the virtual interference signal data represented by the magnetic field component based on the current component of the interference signal data extracted by the interference signal source extraction unit,
The signal processing apparatus according to claim 4 , wherein the target signal extraction unit extracts the target signal of interest from the measurement data, which is magnetic field data, and the virtual interference signal data. The interference signal source extraction unit is
Information indicating the positional relationship between the signal source of the signal of interest and the signal source of the interfering signal is received, and interfering signal data is extracted based on the received information. 1. The signal processing device according to claim 1. The signal processing according to any one of claims 1 and 3 to 5, wherein the target signal extracting unit extracts the target signal using a signal subspace projection method. Device. The measurement execution unit includes the signal of interest generated by a nerve-induced magnetic field generated by nerve activity of a subject induced by electrical stimulation, and the interference signal generated by an interference magnetic field generated by the electrical stimulation. Get measurement data,
The signal source of the interfering signal exists at a site other than the measurement target site that generates the signal of interest in the subject, and generates an interfering magnetic field based on electrical stimulation for inducing neural activity. 3. The signal processing apparatus according to claim 2, wherein the signal processing apparatus is a site or an electrode attached to the subject for applying the electrical stimulation to the subject . The measurement execution unit generates the signal of interest generated by a nerve-induced magnetic field generated by neural activity in the palm of the subject induced by electrical stimulation received at the fingertip, and an interfering magnetic field generated by the electrical stimulation. The signal processing device according to any one of claims 1 to 8, wherein the measurement data including the interfering signal and the measurement data obtained. The signal processing device according to any one of claims 1 to 9, wherein the measurement execution unit has a plurality of magnetic sensors arranged at positions facing the measurement target site of the subject. . a signal of interest generated by a nerve-induced magnetic field generated by nerve activity of a subject induced by electrical stimulation ; A signal processing method by a signal processing device having a measurement execution unit that acquires measurement data including
estimating a signal source in an extraction target region containing the signal source of the signal of interest and the signal source of the interfering signal based on the measurement data;
extracting interfering signal data generated from the signal source of the interfering signal based on the estimation result of the signal source;
extracting a signal of interest by removing a common portion between the measurement data acquired by the measurement execution unit and the interfering signal data;
The signal source of the interfering signal exists at a site other than the measurement target site that generates the signal of interest in the subject, and generates an interfering magnetic field based on electrical stimulation for inducing neural activity. site or electrode attached to the subject for providing the electrical stimulation to the subject
A signal processing method characterized by: The measurement data is magnetic field data,
In estimating a signal source, obtaining a current distribution in the extraction target area, estimating a signal source of the interfering signal based on the obtained current distribution,
12. The method according to claim 11, wherein in extracting the interference signal data, the interference signal data is extracted by regarding a current component of the interference signal data generated from the estimated signal source of the interference signal as a magnetic field component. Signal processing method. generating virtual jamming signal data from the extracted jamming signal data;
11. The method of extracting a signal of interest, wherein the virtual interfering signal data is used instead of the interfering signal data to remove a common portion between the measurement data and the virtual interfering signal data. 13. The signal processing method according to 12. In generating the virtual interference signal data, generating the virtual interference signal data indicated by a magnetic field component based on the current component of the extracted interference signal data;
14. The signal processing method according to claim 13, wherein, in extracting the signal of interest, the signal of interest is extracted from the measurement data, which is magnetic field data, and the virtual interference signal data. a signal of interest generated by a nerve-induced magnetic field generated by nerve activity of a subject induced by electrical stimulation ; A signal processing program that causes a signal processing device having a measurement execution unit that acquires measurement data including
In the signal processing device,
estimating a signal source in an extraction target region including the signal source of the signal of interest and the signal source of the interfering signal;
extracting interference signal data generated from the signal source of the interference signal based on the estimation result of the signal source based on the measurement data;
Extracting a signal of interest by removing a common portion between the measurement data acquired by the measurement execution unit and the interfering signal data ,
The signal source of the interfering signal exists at a site other than the measurement target site that generates the signal of interest in the subject, and generates an interfering magnetic field based on electrical stimulation for inducing neural activity. site or electrode attached to the subject for providing the electrical stimulation to the subject
A signal processing program characterized by: The measurement data is magnetic field data,
In the signal processing device,
In estimating a signal source, obtaining a current distribution in the extraction target region, and estimating a signal source of the interfering signal based on the obtained current distribution,
In extracting the interference signal data, extracting the interference signal data by regarding a current component of the interference signal data generated from the estimated signal source of the interference signal as a magnetic field component.
16. The signal processing program according to claim 15, characterized by : In the signal processing device,
generating virtual interfering signal data from the extracted interfering signal data;
15. The method according to claim 15, wherein, in extracting a signal of interest, the virtual interfering signal data is used instead of the interfering signal data to remove a common portion between the measurement data and the virtual interfering signal data. 17. The signal processing program according to 16. In the signal processing device,
generating the virtual interference signal data represented by the magnetic field component based on the current component of the extracted interference signal data in generating the virtual interference signal data;
18. The signal processing program according to claim 17, wherein, in extracting the signal of interest, the signal of interest is extracted from the measurement data, which is magnetic field data, and the virtual interference signal data.

Images