WO2024075669A1 - Method and device for suggesting positioning of electromagnetic sensors, and current source location estimation method - Google Patents

Abstract

This method for suggesting the positioning of electromagnetic sensors involves suggesting the positioning of a plurality of electromagnetic sensors positioned on the surface of a living body around a target region inside a subject, using a forward model for estimating electromagnetic fields produced at the positions of the plurality of electromagnetic sensors by biological activity inside the subject, the method including: a step for setting calculation conditions including the position of the target region and the positions at which the plurality of electromagnetic sensors can be positioned; a step for calculating, on the basis of the calculation conditions that have been set, suggested positioning for all or some of the plurality of electromagnetic sensors, which are for estimating the position of a current source inside the subject; and a step for outputting the suggested positioning.

Description

Electromagnetic sensor layout proposal method and layout proposal device, and current source position estimation method

This disclosure relates to a method and device for proposing placement of electromagnetic sensors, and a method for estimating the position of a current source.

A technology is known that uses multiple electromagnetic sensors placed on the surface of a subject's body to detect biological activity inside the body. For example, in magnetoencephalography (MEG), multiple sensors placed on the subject's head can detect brain magnetism that occurs with neural activity. In order to clarify brain function, it is desirable to accurately estimate the location of the brain's active area, which acts as a current source (hereinafter referred to as "current source estimation"). One measure to improve the accuracy of current source estimation is to measure the brain magnetism with high sensitivity. To measure the brain magnetism with high sensitivity, the positional relationship between the sensor and the head is important. Specifically, it is important to bring the sensor close to the brain's active area and to place the sensor in an appropriate position according to brain activity.

Conventional magnetoencephalography uses multiple superconducting quantum interference device (SQUID) sensors, each of which requires cooling with liquid helium. Each SQUID sensor is placed in a dewar filled with liquid helium, and the placement of each SQUID is fixed.

On the other hand, there has been active development of small magnetic sensors that operate at room temperature, such as OPM (Optically Pumped Magnetometer) sensors, TMR (Tunnel Magneto Resistance) sensors, and NVC (Nitrogen-Vacancy Center in Diamond) sensors. These sensors do not require cooling with liquid helium and can be easily repositioned, making it easy to change the position of the sensor according to the location of brain activity, which was previously difficult. For example, JP 2020-151023 A discloses identifying the location of the Brodmann's area of the subject's head from an MRI (Magnetic Resonance Imaging) image of the subject's head, and placing one OPM sensor on the surface of the head close to the identified Brodmann's area.

JP 2020-151023 A

Even when current source estimation is performed using a small electromagnetic sensor that operates at room temperature, such as an OPM sensor, it is desirable to place multiple sensors on the surface of the living body, rather than just one, as with conventional SQUIDs. Furthermore, if the number of sensors is limited for some reason, it is desirable to place the limited number of sensors in appropriate positions. However, JP 2020-151023 A does not mention anything about the appropriate placement of multiple sensors.

Also, even when the placement of each sensor is fixed, as in the case of SQUID, there has been no sufficient consideration given to which of the outputs of multiple sensors should be used.

The present disclosure has been made to solve the above problems, and one objective of the present disclosure is to propose an arrangement of multiple sensors suitable for current source estimation. Another objective of the present disclosure is to provide a method for optimally utilizing the outputs of multiple sensors.

The method for proposing the placement of electromagnetic sensors according to the present disclosure is a method for proposing the placement of multiple electromagnetic sensors to be placed on the biological surface surrounding a target area inside the subject, using a forward model for estimating an electromagnetic field generated at the positions of the multiple electromagnetic sensors due to biological activity inside the subject, and includes the steps of setting calculation conditions including the position of the target area and possible positions for placing the multiple electromagnetic sensors, calculating a proposed placement of all or a part of the multiple electromagnetic sensors to estimate the position of a current source inside the subject based on the set calculation conditions, and outputting the proposed placement.

The current source position estimation method disclosed herein is a method for estimating the position of a current source inside a subject using the detection results of multiple electromagnetic sensors that are placed on the biological surface of the subject and each detects an electromagnetic field generated by biological activity taking place in a target area inside the subject, and includes the steps of setting calculation conditions including the position of the target area and the arrangement of the multiple electromagnetic sensors, calculating priorities of all or a part of the multiple electromagnetic sensors based on the calculation conditions, and estimating the position of the current source using the detection results of the electromagnetic sensors selected based on the priorities.

The electromagnetic sensor placement suggestion device according to the present disclosure is a device that proposes the placement of multiple electromagnetic sensors to be placed on the biological surface surrounding a target area inside the subject, using a forward model for estimating the electromagnetic field generated at the positions of the multiple electromagnetic sensors due to biological activity inside the subject, and includes an input unit to which calculation conditions including the position of the target area and possible positions of the multiple electromagnetic sensors are input, a calculation device that calculates a proposed placement of all or a part of the multiple electromagnetic sensors to estimate the position of a current source inside the subject based on the calculation conditions, and an output unit that outputs the proposed placement.

The layout proposal method and layout proposal device disclosed herein can propose a layout of multiple sensors suitable for current source estimation.

The position estimation method disclosed herein allows the position of a current source to be accurately estimated using the detection results of a high-priority sensor selected from among multiple electromagnetic sensors.

1 is a diagram showing a schematic diagram of an overall configuration of a data processing system used in a method for proposing an arrangement of electromagnetic sensors; FIG. 13 is a diagram conceptually illustrating a method for proposing an arrangement of an OPM sensor. 13 is a flowchart illustrating an example of a procedure for a placement proposal process. FIG. 13 is a diagram showing an example of sensor arrangement and measurement results when magnetoencephalography is performed twice using 11 OPM sensors. FIG. 5 is a diagram comparing the measurement accuracy of the first measurement (uniform arrangement) shown in FIG. 4 with the measurement accuracy of the second measurement (optimal arrangement).

Below, the embodiments of the present disclosure will be described in detail with reference to the drawings. Note that the same or equivalent parts in the drawings will be given the same reference numerals and their description will not be repeated.

[First embodiment]
<System Configuration>
1 is a diagram showing a schematic diagram of an overall configuration of a data processing system 1 used in a method for proposing an arrangement of an electromagnetic sensor according to the present embodiment. In this embodiment, a case where an OPM sensor 2, which is a small magnetic sensor that operates at room temperature, is used as an example of an electromagnetic sensor will be described.

The data processing system 1 includes a plurality of OPM sensors 2, an input device 3, an output device 4, and a processing device 10. The data processing system 1 is configured to estimate the location of biological activity (current source) in the subject's brain using the detection results from the plurality of OPM sensors 2.

Each of the multiple OPM sensors 2 is placed on the surface of the subject's head by a user (medical personnel, etc.) and detects the magnetic field generated at each placement position by biological activity taking place in the subject's brain. The multiple OPM sensors 2 may be in direct contact with the surface of the subject's head, or may not be in direct contact with the surface of the subject's head. For example, a cover member that holds the sensor 2 may be attached to the subject's head. Note that FIG. 1 shows an example in which four OPM sensors 2 are placed on the surface of the subject's head.

Since each OPM sensor 2 operates at room temperature, it does not need to be placed inside a Dewar like the SQUID described above. Therefore, the placement of multiple OPM sensors 2 can be easily changed depending on brain activity.

The input device 3 is, for example, a pointing device such as a keyboard or a mouse, and accepts input information (such as calculation conditions described below) from the user. The information input to the input device 3 is sent to the processing device 10. The output device 4 is, for example, a liquid crystal display (LCD) panel, and is a display that shows information to the user.

The processing device 10 has, as its main hardware elements, a sensor interface 11, an input interface 12, an output interface 13, a storage device 14, and an arithmetic unit 15. The processing device 10 may be realized, for example, by a general-purpose computer, or may be realized by a computer (such as a server) dedicated to the data processing system 1.

The sensor interface 11 is an interface for connecting the processing device 10 to the multiple OPM sensors 2, and realizes the input and output of signals between the processing device 10 and the multiple OPM sensors 2. The input interface 12 is an interface for connecting the processing device 10 to the input device 3, and realizes the input and output of signals between the processing device 10 and the input device 3. The output interface 13 is an interface for connecting the processing device 10 to the output device 4, and realizes the input and output of data between the processing device 10 and the output device 4.

The storage device 14 stores information (programs, etc.) used for processing by the arithmetic device 15. Note that input information (such as calculation conditions described below) input by the user to the input device 3 is stored in the storage device 14.

The calculation device 15 has a CPU (Central Processing Unit) and performs "current source estimation" to estimate the position of the current source (place of biological activity) in the subject's brain using the information stored in the storage device 14 and the measurement results of multiple OPM sensors 2.

The computing device 15 causes the output device 4 to display the position of the current source obtained by current source estimation as the location of biological activity. By looking at the content displayed on the output device 4, the user can grasp the location of the location of biological activity in the subject's brain.

<Proposal of sensor arrangement suitable for current source estimation>
In order to accurately estimate the position of a current source in the brain, it is desirable to measure the brain magnetism with high sensitivity using multiple OPM sensors 2. In order to measure the brain magnetism with high sensitivity, the positional relationship between the multiple OPM sensors 2 and the current source in the brain is important. Specifically, it is important to place the multiple OPM sensors 2 in appropriate positions according to the position of the current source in the brain.

The calculation device 15 according to this embodiment is configured to perform a process for proposing an arrangement of the OPM sensor 2 suitable for current source estimation (hereinafter also referred to as the "arrangement proposal process"). Hereinafter, the arrangement of the OPM sensor 2 proposed by the arrangement proposal process is also simply referred to as the "proposed arrangement."

FIG. 2 is a diagram conceptually showing a method for proposing the placement of the OPM sensor 2 by the placement proposal process.
In the placement proposal process, the steps of setting calculation conditions, calculating the proposed placement, and outputting the proposed placement are carried out in this order.

[Calculation condition settings]
In the step of setting the calculation conditions, the calculation conditions used for calculating the proposed arrangement are set based on the input information input to the input device 3 and the information stored in the storage device 14. The calculation conditions include "possible sensor arrangement positions,""number of sensors s,""sensor SNR (Signal-to-Noise Ratio),""brainmodel,""targetregion,""objective function f," and the like.

"Sensor placement positions" is data indicating positions where OPM sensors 2 can be placed on the surface of the subject's head. Specifically, "sensor placement positions" includes the maximum number of OPM sensors 2 that can be placed (hereinafter also referred to as the "maximum number m") and the coordinates of the positions where the sensors can be placed. The coordinates of the sensor placement positions can be determined, for example, based on image data obtained by MRI or a three-dimensional scanner of the subject's head.

Note that while FIG. 2 shows an example in which the coordinates of the possible sensor placement positions are discrete values, it is also possible to treat the coordinates of the possible sensor placement positions as continuous values by defining the surface coordinates of the head in a form similar to an equation for a sphere or a plane. In addition, the coordinates of the possible sensor placement positions may be virtual positions where the OPM sensor 2 cannot actually be placed.

The "number of sensors s" is the number of OPM sensors 2 actually placed on the subject's head surface. The number of sensors s is a value equal to or less than the maximum number of sensors m. In the example shown in FIG. 2, the number of sensors s is "4."

The "sensor SNR" is the SNR of the OPM sensor 2. The "sensor SNR" is used to determine the "constant λ" used in the placement calculation described below.

The "brain model" is a model of the subject's brain structure using a triangular mesh or the like. The brain model may be one that is generated based on the results of measuring the subject's brain by MRI, or it may be one that assumes the subject's brain is a standard brain and substitutes it with an existing standard brain model.

The "target region" is a location within the brain model where biological activity is expected to occur. In the calculation of the proposed layout described below, the vertices within the target region within the brain model will be treated as the positions of the current source (current dipole). The number of vertices included in the target region may be one or multiple.

The "objective function f" is a function used in the calculation of the proposed placement, which will be described later. As will be described later, the "objective function f" is determined according to the "target area."

In this embodiment, the "objective function f" is set to a function that maximizes the diagonal sum (sum of diagonal components) of the vertices in the target region, but this is not limited to this. Any of the functions listed in Table 1 of Olaf Hauk et. al., "Towards an objective evaluation of EEGMEG source estimation methods - The linear approach (Olaf Hauk)" may be used as the objective function.

[Calculation of proposed placement]
In the step of calculating the proposed layout, a sensor layout suitable for the set calculation conditions is derived as the proposed layout by theoretical calculation. In this embodiment, each proposed layout is assigned a "priority p" indicating the order of importance. The calculation method of the proposed layout will be described in detail later.

[Output of proposed layout]
In the step of outputting the proposed layout, data of the proposed layout obtained from the theoretical calculation is output. For example, when the data of the proposed layout is output to the output device 4, the proposed layout is displayed on the output device 4.

In the example shown in Figure 2, the proposed layouts are displayed by priority order. Specifically, the most important layout is shown as "priority order p=1", with the layout with priority order 1 displayed above it. The second most important layout is shown as "priority order p=2", with the layouts with priorities 1 and 2 displayed above it. The third most important layout is shown as "priority order p=3", with the layouts with priorities 1 to 3 displayed above it. The fourth most important layout is shown as "priority order p=4", with the layouts with priorities 1 to 4 displayed above it.

[Procedure for placement proposal processing]
FIG. 3 is a flowchart showing an example of the procedure of the above-mentioned placement proposal process.

First, the calculation device 15 sets calculation conditions based on the input information input to the input device 3 and the information stored in the storage device 14 (step S10). As described above, the calculation conditions include possible sensor placement positions (maximum number of placements m, coordinates of each placement), number of sensors s, sensor SNR, brain model, target region, objective function f, etc.

Then, the calculation device 15 sets the priority p to the initial value "1" (step S11). After that, the calculation device 15 moves the process to step S30.

Then, the computing device 15 arbitrarily selects one sensor position whose priority has not yet been determined from among the multiple possible sensor placement positions (step S30).

Then, the calculation device 15 calculates the forward model matrix G (step S31). The forward model matrix G is a matrix that indicates a forward model solution that expresses the correspondence between the current dipole moment of a current source when a current dipole moment is applied to the vertices of the brain model to make them current sources, and the magnetic field generated at the position of the OPM sensor 2 (the magnetic field measured by the OPM sensor 2). In other words, the forward model matrix G satisfies the following formula (1).

In equation (1), "B" is the magnetic field vector generated at the position of the OPM sensor 2, "J" is the density vector of the true current of the current source in the brain, and "G" is the forward model matrix (a matrix that indicates the forward model solution). The forward model matrix G is a matrix that converts the current of the current source into a magnetic field generated at the position of the OPM sensor 2, and is also called the "lead field matrix" in the field of magnetoencephalography. This forward model matrix G corresponds to an example of the "forward model" of this disclosure.

The calculation unit 15 calculates the forward model matrix G _p of the priority level p by using the following formula (2).

Specifically, the calculation device 15 calculates a forward model matrix G p of the priority level p by adding a forward model matrix g corresponding to one sensor whose priority level is undetermined and selected in step S30 to the forward model matrix G p- ₁ of the previous priority level p _-1 .

The size of the forward model matrix _Gp is (priority p) × (number of vertices v). Therefore, when the priority p reaches the number of sensors s, the forward model matrix _Gs becomes (number of sensors s) × (number of vertices v). The number of vertices v is the number of vertices of the brain model (the number of points where current dipoles simulating neural activity are placed). When determining the position of priority p=1, there is no sensor position with a determined priority yet, and the forward model matrix _G0 is an empty matrix. Therefore, the forward model matrix _G1 of priority 1 becomes _G1 =g.

Then, the calculation device 15 uses the forward model matrix G calculated in step S31 to calculate the estimated model matrix R (step S32). The estimated model matrix R is a matrix corresponding to an estimated model for estimating the position of a current source in the brain from the detection results of the OPM sensor 2 placed on the head surface. The estimated model matrix R satisfies the following equation (3).

In equation (3), "J" is the density vector of the true current of the current source in the brain, "J'" is the density vector of the estimated current of the current source in the brain, and "R" is the estimated model matrix (a matrix that indicates the estimated model solution). The estimated model matrix R is a matrix that converts the true current (true value) of the current source into a magnetic field generated at the sensor position, and further converts the converted magnetic field back into an estimated current (estimated solution) of the current source. The estimated model matrix R is also called a "resolution matrix" in the field of magnetoencephalography. This estimated model matrix R corresponds to an example of an "estimated model" in this disclosure.

The calculation unit 15 calculates the estimation model matrix R _p of the priority level p by using the following formula (4).

In equation (4), “G _p ′” is a transposed matrix of the forward model matrix G _p , “λ” is a constant determined according to the sensor SNR, and “I” is a unit matrix. Note that the matrix size of the estimated model matrix R _p is (number of vertices v) × (number of vertices v).

Next, the calculation device 15 calculates the difference matrix ΔR _p by using the following equation (5) (step S33).

In equation (5), "R _p " is the estimated model matrix R of the priority order p, and "R _p-1 " is the estimated model matrix R of the priority order p-1.

Next, the calculation device 15 determines whether or not the difference matrix ΔR _p has been calculated for all sensor positions whose priorities have not yet been determined (step S34).

If the difference matrix ΔR _p has not been calculated for all sensor positions whose priorities are undetermined (NO in step S34), the calculation device 15 returns the process to step S30 and repeats the processes of steps S30 to S34 until the difference matrix ΔR _p has been calculated for all sensor positions whose priorities are undetermined.

When the difference matrix ΔR _p has been calculated for all the sensor positions whose priority is not yet determined (YES in step S34), the calculation device 15 identifies the sensor position whose priority is not yet determined and whose objective function f(ΔR _p ) based on the difference matrix ΔR _p is optimal, and sets the identified sensor position as the position whose priority is p (step S40). The objective function f is determined according to the position of the target region, as described later.

Then, the calculation device 15 determines whether the priority p has reached the number of sensors s (step S41). If the priority p has not reached the number of sensors s (NO in step S41), the calculation device 15 moves the process to step S20 and increments the priority p by 1, and then repeats the processes of steps S30 to S34. The processes of steps S30 to S34 are repeated until the priority p reaches the number of sensors s.

By repeating this process, the placement of sensors with priority levels p = 1 to s is determined sequentially in descending order of priority.

For example, assume that the maximum placement number m is "10" and the number of sensors s is "3." In this case, first the position of priority p=1 is determined, then the position of priority p=2 is determined, and finally the position of priority p=3 is determined.

At the stage of determining the position of priority level p=1, the priorities of all 10 possible placement positions have not yet been determined, so the forward model matrix G ₁ , the estimated model matrix R ₁ and the difference matrix ΔR ₁ for determining the position of priority level p=1 are calculated separately for each of the 10 possible placement positions. Then, the position at which the objective function f based on the difference matrix ΔR ₁ is optimal is set as the position of priority level p=1.

When the position of the priority p=1 is set, a process of determining the position of the next priority p=2 is performed. At this stage, the forward model matrix G ₂ , the estimated model matrix R ₂ and the difference matrix ΔR ₂ are calculated separately for the remaining nine possible positions excluding the position of the priority p=1. At this time, G ₂ , R ₂ and ΔR ₂ are calculated using G ₁ , R ₁ and ΔR ₁ for determining the position of the previous priority p=1. Therefore, G ₂ , R ₂ and ΔR ₂ are values that reflect the influence when a sensor is placed at the position of the priority p=1. Then, the position where the objective function f based on ΔR ₂ is optimal is set as the position of the priority p=2. Therefore, the position of the priority p=2 is set by taking into account the position of the previous priority p=1.

When the position of priority p=2 is set, a process of determining the position of the final priority p=3 is performed. At this stage, the forward model matrix G ₃ , the estimated model matrix R ₃ and the difference matrix ΔR ₃ are calculated separately for the remaining eight possible placement positions excluding the positions of priority p=1, 2. At this time, G ₃ , R ₃ and ΔR ₃ are calculated using G ₂ , R ₂ and ΔR ₂ of the immediately previous priority p=2. G ₂ , R ₂ and ΔR ₂ are further calculated using G ₁ , R ₁ and ΔR ₁ of the immediately previous priority p=1. Therefore, G ₃ , R ₃ and ΔR ₃ are values that reflect the influence of the positions of priority p=1, 2. Then, the position where the objective function f based on ΔR ₃ is optimal is set to the position of priority p=3. Therefore, the position of priority p=3 is set taking into account the positions of priorities p=1, 2 up to the immediately previous priority.

When the priority p reaches the number of sensors s (YES in step S41), the calculation device 15 outputs the sensor placements with priority p=1 to s as proposed placements (step S50). For example, the output device 4 displays each of the proposed placements with priority p=1 to s.

By performing the above-described placement suggestion process, the processing device 10 can provide the user with a proposed placement of the s OPM sensors 2 (a sensor placement suitable for estimating the position of the current source in the subject's brain). The user can accurately estimate the position of the current source by placing the s OPM sensors 2 according to the proposed placement displayed on the output device 4. If the proposed placement includes a virtual position where the OPM sensor 2 cannot actually be placed, the OPM sensor 2 can be placed in a position closest to the virtual position and where the OPM sensor 2 can actually be placed.

[Calculation of forward model matrix G]
A detailed description will be given below of a method for calculating the forward model matrix G. As described above, the forward model matrix G is a matrix that satisfies "B = GJ" shown in equation (1). Specifically, the forward model matrix G is a matrix that converts the current density vector "J" of a current source in the brain into a magnetic field vector "B" generated at the position of the OPM sensor 2.

The relationship between the current generated by the current source in the brain and the magnetic field generated at the position of the OPM sensor 2 can be expressed by the following equation (6) using Biot-Savart's law.

In equation (6), "B" is the magnetic field vector generated at the position of the OPM sensor 2, "r" is the position (coordinate) of the OPM sensor 2, "J" is the current density vector of the current source in the brain, and "r'" is the position (coordinate) of the current source in the brain.

The components of the forward model matrix G can be found by calculating equation (6) above. "J" is a current density vector, and if its direction is not specified, a huge number of equations will have to be calculated; however, in the field of current source estimation, it is considered reasonable to assume that the current flows in a direction perpendicular to the brain model, so "J" can be uniquely determined. The position "r" of the OPM sensor 2 can be determined by the possible sensor placement positions set in the calculation conditions. The position "r'" of the current source can be determined by the vertex position of the brain model set in the calculation conditions. For this reason, by calculating the Biot-Savart equation for one current source (vertex) for the maximum number of sensor placements m, the components of the forward model matrix G for that one current source can be calculated.

Specifically, Biot-Savart's law can be written as the maximum number m of sensors per current source, as shown in equation (7) below.

Here, by taking the magnitude of the current from the current source as unit current, it is possible to determine the magnetic field B generated at the position of the OPM sensor 2 per unit current.

Furthermore, to make it easier to understand, consider the case where the OPM sensor 2 has sensitivity in only one axial direction. In this case, when the components of each matrix are visualized, "B = GJ" in the above formula (1) becomes the following formula (8).

The calculation device 15 can calculate the components of the forward model matrix G using the above-mentioned method.

[Relationship between target area and objective function f]
The "target region" set in the calculation conditions is used to set the "objective function f". The relationship between the target region and the objective function f will be described in detail below. As described above, the objective function f is a function based on the difference matrix ΔR. However, since the difference matrix ΔR is obtained by taking the difference of the estimated model matrix R, the objective function f will be described here as being based on the estimated model matrix R for ease of understanding.

As described above, the estimated model matrix R is a matrix that satisfies "J' = RJ" shown in equation (3). Specifically, the estimated model matrix R is a matrix that converts the true current J into the estimated current J'.

If we visualize each matrix element in equation (3), it becomes as shown in equation (9) below.

In equation (9), "j ₁ to j _n " are true currents at each vertex of the brain model, and "j ₁ ' to j _n '" are estimated currents at each vertex of the brain model.

If the target region is an area including the four vertices of the brain model (vertices 1, 2, 3, and 4), then the first to fourth rows and first to fourth columns of the estimated model matrix R are the matrix elements related to the target region, as shown in equation (10) below.

Looking at the estimation model matrix R in detail, it is possible to see the influence that the true current of vertex 1 has on the estimated current of vertices 1 to v, and the influence that the true current of vertices 1 to v has on the estimated current of vertex 1. Taking these into consideration, one ideal form of the estimation model matrix R is a matrix that has values only in the diagonal components where true values exist. In other words, the objective function f can be set to a function that maximizes the diagonal sum (the sum of the diagonal components) of the vertices in the target region. In view of this, the calculation device 15 determines the objective function f according to the position of the target region (more specifically, the vertices in the target region).

If the target region to be optimized is taken broadly and includes vertices 1 to 4, the objective function f can be set to a function that maximizes the diagonal sum of elements r11 to r44 of the estimated model matrix R, as shown in equation (11) below. In situations where the region where brain activity occurs can be predicted but the exact location cannot be narrowed down, it is conceivable to take a broad target region like this.

On the other hand, if the target region to be optimized is narrowed to include only vertex 1, the objective function f can be set to a function that maximizes the diagonal component r11 of the estimated model matrix R. In situations where the region in which brain activity will occur can be predicted with pinpoint precision, it is expected that the target region will be narrowed in this way.

The objective function f may also be set to a function that maximizes the ratio between the diagonal sum (sum of diagonal components) and the off-diagonal sum (sum of off-diagonal components) of the target region, as shown in the following formula (12). Alternatively, the objective function f may be set to a function that maximizes the ratio between the diagonal sum of the target region and other elements.

As described above, the calculation device 15 according to this embodiment calculates the forward model matrix G and the estimated model matrix R based on calculation conditions including the target area and possible sensor placement positions. The calculation device 15 then sets and outputs a proposed placement (priority order p) of the OPM sensors 2 using as an index the optimum objective function f set based on the estimated model matrix R. By placing multiple OPM sensors 2 according to the proposed placement, the user can efficiently perform highly sensitive magnetoencephalography.

The layout proposal method according to this embodiment can propose a layout of multiple OPM sensors 2 suitable for current source estimation. This makes it possible to efficiently optimize the sensor layout, which was difficult with conventional SQUIDs. Furthermore, it enables highly sensitive (highly accurate) magnetoencephalography. It also makes it possible to reduce the number of sensors required for current source estimation, thereby reducing the hardware costs of the magnetoencephalography system.

[Embodiment 2]
In the second embodiment, in order to efficiently perform magnetoencephalography with higher accuracy using a limited number of OPM sensors 2, the magnetoencephalography is performed in two steps as described below.

Figure 4 shows an example of sensor arrangement and measurement results when performing two magnetoencephalographic measurements using 11 OPM sensors 2.

In the first measurement, since it is unclear where in the brain the current source (the true abnormal area) is located, the entire brain is treated as the test area and 11 OPM sensors 2 are evenly placed on the surface of the head. The position of the current source is then tentatively determined from the measurement results of the 11 evenly placed OPM sensors 2. In the example shown in Figure 4, since the first measurement results show more measurement results with stronger intensity in the left brain than in the right brain, the left brain is tentatively determined as the tentative position of the current source. Note that the part of the left brain where intensity above a predetermined value is concentrated may also be determined as the tentative position of the current source.

In the second measurement, the provisional position of the current source tentatively measured in the first measurement (left brain) is set as the "target area" and the above-mentioned placement proposal process is performed, and the 11 OPM sensors 2 are repositioned according to the proposed placement obtained in the placement proposal process. The position of the current source is then officially measured from the measurement results of the 11 repositioned OPM sensors 2. In the second measurement, by repositioning the OPM sensors 2 to the optimal position based on the measurement results of the first measurement, the measurement result of maximum intensity is seen in a position closer to the true value (true abnormal area). In other words, the estimation accuracy of the current source has improved in the second measurement.

Fig. 5 is a diagram comparing the measurement accuracy of the first measurement (uniform arrangement) and the second measurement (optimal arrangement) shown in Fig. 4. It can be said that the estimation accuracy of the current source is higher when the position error between the true value and the estimated value (i.e., the shortest distance between the coordinate _j1 of the maximum intensity in the true value and the coordinate _j2 of the maximum intensity in the estimated current) is smaller and when the spatial spread of the estimated current (the maximum length of the area where the estimated current is measured) is smaller.

As shown in Figure 5, in the first measurement (even placement), the position error between the true value and the estimated value was 33.16 mm, and the spatial spread of the estimated current was 48.28 mm, whereas in the second measurement (optimal placement), the position error between the true value and the estimated value was 5.18 mm, and the spatial spread of the estimated current was 34.38 mm, which is an improvement. In particular, the position error between the true value and the estimated value has improved significantly (by about 28 mm), and with optimal placement, the maximum intensity of the estimated current is only about 5 mm away from the true value (the abnormal region that is the true source of activity).

In this way, in the first measurement in the second embodiment, multiple OPM sensors 2 are evenly arranged on the head surface to tentatively measure the position of the current source, the tentatively measured position of the current source is set as the target region and a placement proposal process is performed, and the multiple OPM sensors 2 are re-arranged according to the proposed placement obtained in the placement proposal process to determine the position of the current source for the second measurement. This makes it possible to efficiently perform high-precision magnetoencephalography using a limited number of OPM sensors 2.

In the first provisional measurement, it is sufficient to measure the provisional position of the current source, and it is not necessarily limited to using the OPM sensor 2. For example, in the first provisional measurement, the provisional position of the current source may be measured using fMRI (functional Magnetic Resonance Imaging) or the like. In addition, a doctor's opinion may be involved in determining the provisional position of the current source.

[Modification 1]
For example, in step S50 of FIG. 3 described above, in addition to displaying the proposed layouts of each priority level on the output device 4, the position error between the target area and the result of current source estimation (the position error between the true current and the estimated current) for each proposed layout may be displayed on the output device 4.

In this way, the user can look at the output device 4 and check whether the position error in the proposed layout for each priority order is within an acceptable range, and then decide the number of sensors s to be used for current source estimation. Therefore, it is also possible to reduce the number of sensors s to be used for current source estimation as necessary, making it possible to reduce the hardware costs of the magnetoencephalography system.

[Modification 2]
The sensor to which the layout proposal method according to the present disclosure can be applied is not limited to the OPM sensor 2. The layout proposal method according to the present disclosure can also be applied to, for example, a fluxgate sensor, a magneto resistance (MR) sensor, a magneto impedance (MI) sensor, a coil type sensor, or a nitrogen-vacancy center in diamond (NVC) sensor instead of or in addition to the OPM sensor 2.

[Modification 3]
The field to which the placement proposal method according to the present disclosure can be applied is not limited to the field of magnetoencephalography. The placement proposal method according to the present disclosure can also be applied to fields in which biomagnetic fields other than those of the brain (such as magnetic fields generated by biological activity in the heart, spinal cord, peripheral nerves, or muscles) are measured.

[Modification 4]
The sensors to which the layout proposal method according to the present disclosure can be applied are not limited to magnetic sensors. That is, the layout proposal method according to the present disclosure can also be applied to potential sensors that detect potentials generated by electrical activity of a living body, such as an electroencephalogram (EEG) sensor or an electromyogram (EMG) sensor.

[Modification 5]
The placement proposal method according to the present disclosure can also be applied to SQUIDs whose positions are difficult to change. As described above, the SQUIDs are placed in a dewar filled with liquid helium, and it is difficult to easily change their positions. However, it is possible to place a large number of SQUIDs (e.g., tens or hundreds) on the subject's head in advance. The placement proposal method described above can be applied to a process of proposing the positions of SQUIDs useful for current source estimation among the large number of SQUIDs placed in advance on the subject's head.

Specifically, multiple SQUIDs may be placed on the subject's head in advance, calculation conditions may be set including the position of the target area and the placement of the multiple SQUIDs, priorities of all or some of the multiple SQUIDs may be calculated based on the calculation conditions, and the position of the current source in the subject's brain may be estimated using the detection results of the electromagnetic sensor selected based on the priorities.

In this way, a high-priority SQUID can be selected from among the many SQUIDs pre-positioned on the subject's head, and the current source can be estimated using the measurement results of the selected SQUID. Therefore, even when using a SQUID whose position is difficult to change, it is possible to perform accurate current source estimation by pre-positioning a large number of electromagnetic sensors on the subject's head.

[Aspects]
It will be understood by those skilled in the art that the above-described embodiment and its modifications are specific examples of the following aspects.

(Section 1) A method for proposing the placement of electromagnetic sensors according to one embodiment is a method for proposing the placement of multiple electromagnetic sensors to be placed on a biological surface surrounding a target area inside a subject, using a forward model for estimating an electromagnetic field generated at the positions of the multiple electromagnetic sensors due to biological activity inside the subject, and includes the steps of setting calculation conditions including the position of the target area and possible positions for placing the multiple electromagnetic sensors, calculating a proposed placement of all or a part of the multiple electromagnetic sensors to estimate the position of a current source inside the subject based on the set calculation conditions, and outputting the proposed placement.

According to the placement suggestion method described in paragraph 1, a proposed placement of sensors is calculated and output based on the calculation conditions. This makes it possible to provide the user with a proposed placement of sensors suitable for current source estimation. The user can accurately estimate the position of the current source by placing multiple electromagnetic sensors according to the proposed placement.

(2) In the method for proposing placement of an electromagnetic sensor described in 1, the step of calculating the proposed placement includes a step of calculating the priority of all or a part of the possible placement positions.

The placement suggestion method described in paragraph 2 can provide the user with the priorities of all or part of the possible placement locations.

(3) The method for proposing an arrangement of an electromagnetic sensor described in 1 or 2 further includes a step of tentatively measuring the position of a current source. The step of setting the calculation conditions includes a step of setting the tentatively measured position of the current source as the position of the target region.

According to the placement proposal method described in Section 3, the positions of the current sources are tentatively measured before the proposed placement is calculated, and the proposed placement is calculated based on the tentatively measured positions of the current sources. This allows the proposed placement to be calculated efficiently with high accuracy.

(4) In the method for proposing an arrangement of an electromagnetic sensor described in any one of paragraphs 1 to 3, the step of setting the calculation conditions includes a step of setting the parameters of the forward model as the calculation conditions.

According to the placement proposal method described in Section 4, the proposed placement can be calculated by setting the parameters of the forward model as the calculation conditions.

(5) In the method for proposing an arrangement of an electromagnetic sensor described in any one of 1 to 4, the step of calculating the proposed arrangement includes a step of calculating the proposed arrangement using an estimation model for estimating the position of a current source from the detection results of all or a part of the multiple electromagnetic sensors. The step of calculating the proposed arrangement includes a step of calculating the proposed arrangement so that the difference between the estimation result of the estimation model obtained by setting the parameters of the estimation model as the calculation conditions and the position of the target area is small.

According to the layout proposal method described in Section 5, the proposed layout can be calculated so that the difference between the estimation result of the estimation model (position of the estimated current) and the position of the target area (position of the true current) is small.

(Section 6) In the method for proposing an arrangement of an electromagnetic sensor described in Section 5, the step of outputting the proposed arrangement includes a step of presenting to a user the difference between the estimation result of the estimation model and the position of the target area in the proposed arrangement.

According to the placement proposal method described in Section 6, the user can confirm the difference between the estimation result of the estimation model and the position of the target region as the accuracy of the current source estimation in the proposed placement.

(7) In the method for proposing placement of an electromagnetic sensor described in any one of 1 to 6, the calculation conditions include at least one of the number of multiple electromagnetic sensors and the SNR (signal-noise ratio) of the multiple electromagnetic sensors in addition to the position of the target area and the possible placement positions.

According to the placement proposal method described in Section 7, the proposed placement can be calculated taking into account at least one of the number of electromagnetic sensors and the SNR.

(8) In the method for proposing an arrangement of an electromagnetic sensor described in any one of 1 to 7, the electromagnetic sensor is an OPM (Optically Pumped Magnetometer) sensor, a fluxgate sensor, an MR (Magneto Resistance) sensor, an MI (Magneto Impedance) sensor, a coil-type sensor, or an NVC (Nitrogen-Vacancy Center in Diamond) sensor.

According to the placement suggestion method described in Section 8, a suggested placement of an OPM sensor, fluxgate sensor, MR sensor, MI sensor, coil-type sensor, or NVC sensor can be provided to a user.

(9) In the method for proposing an electromagnetic sensor placement described in any one of 1 to 8, the electromagnetic sensor is a magnetic sensor or an electric potential sensor that detects a magnetic field generated by electrical activity in the brain, heart, spinal cord, peripheral nerves, or muscles.

The placement suggestion method described in paragraph 9 can provide a user with suggested placements for magnetic sensors or electric potential sensors that detect magnetic fields generated by electrical activity in the brain, heart, spinal cord, peripheral nerves, or muscles.

(10) In the method for proposing an arrangement of an electromagnetic sensor described in any one of paragraphs 1 to 9, the step of outputting the proposed arrangement includes a step of presenting the proposed arrangement to a user.

According to the method for proposing an arrangement described in the tenth aspect, the proposed arrangement can be presented to the user.
(Clause 11) A method for estimating the position of a current source according to one embodiment is a method for estimating the position of a current source inside a subject using detection results of a plurality of electromagnetic sensors arranged on a biological surface of the subject, each of which detects an electromagnetic field generated by biological activity taking place in a target area inside the subject, and includes the steps of setting calculation conditions including the position of the target area and the arrangement of the plurality of electromagnetic sensors, calculating priorities of all or a part of the plurality of electromagnetic sensors based on the calculation conditions, and estimating the position of the current source using the detection results of the electromagnetic sensors selected based on the priorities.

According to the placement proposal method described in Section 11, current source estimation is performed using the detection results of high-priority electromagnetic sensors among a large number of electromagnetic sensors. Therefore, even when using electromagnetic sensors (such as SQUIDs) whose positions are difficult to change, by placing a large number of electromagnetic sensors on the biological surface of the subject in advance, current source estimation can be performed with high accuracy using the detection results of high-priority sensors selected from a large number of electromagnetic sensors.

(Clause 12) An electromagnetic sensor placement suggestion device according to one embodiment is a device that uses a forward model for estimating an electromagnetic field generated at the positions of the electromagnetic sensors due to biological activity inside the subject to suggest a placement of multiple electromagnetic sensors to be placed on a biological surface surrounding a target area inside the subject, and includes an input unit to which calculation conditions including the position of the target area and possible positions of the multiple electromagnetic sensors are input, a calculation device that calculates a proposed placement of all or a part of the multiple electromagnetic sensors to estimate the position of a current source inside the subject based on the calculation conditions, and an output unit that outputs the proposed placement.

The layout proposal method described in paragraph 12 can achieve the same effect as the layout proposal method described in paragraph 1.

The embodiments disclosed herein should be considered in all respects as illustrative and not restrictive. The scope of the present invention is indicated by the claims, not by the description of the embodiments above, and is intended to include all modifications within the meaning and scope of the claims.

1 Data processing system, 2 OPM sensor, 3 Input device, 4 Output device, 10 Processing device, 11 Sensor interface, 12 Input interface, 13 Output interface, 14 Storage device, 15 Arithmetic unit.

Claims (12)

1. A method for proposing an arrangement of a plurality of electromagnetic sensors to be placed on a biological surface around a region of interest within a subject, using a forward model for estimating an electromagnetic field generated at the locations of the plurality of electromagnetic sensors due to biological activity within the subject, the method comprising:
setting a calculation condition including a position of the target area and possible positions of the plurality of electromagnetic sensors;
calculating a proposed arrangement of all or a part of the plurality of electromagnetic sensors for estimating a position of a current source inside the subject based on the set calculation conditions;
and outputting the proposed placement. The method for proposing placement of an electromagnetic sensor according to claim 1, wherein the step of calculating the proposed placement includes a step of calculating priorities of all or a part of the possible placement positions. determining a location of the current source;
3. The method for proposing an arrangement of electromagnetic sensors according to claim 1, wherein the step of setting the calculation conditions includes a step of setting a tentatively measured position of the current source as a position of the target region. The method for proposing an arrangement of an electromagnetic sensor according to claim 1, wherein the step of setting the calculation conditions includes a step of setting parameters of the forward model to the calculation conditions. The step of calculating the proposed arrangement includes a step of calculating the proposed arrangement using an estimation model for estimating a position of the current source from detection results of all or a part of the plurality of electromagnetic sensors;
The method for proposing an arrangement of an electromagnetic sensor as described in claim 1, wherein the step of calculating the proposed arrangement includes a step of calculating the proposed arrangement so that a difference between the estimation result of the estimation model obtained by setting parameters of the estimation model to the calculation conditions and the position of the target area is small. The method for proposing an arrangement of an electromagnetic sensor according to claim 5, wherein the step of outputting the proposed arrangement includes a step of presenting to a user a difference between the estimation result of the estimation model and the position of the target area in the proposed arrangement. The method for proposing placement of electromagnetic sensors according to claim 1, wherein the calculation conditions include at least one of the number of the plurality of electromagnetic sensors and the SNR (Signal-Noise Ratio) of the plurality of electromagnetic sensors in addition to the position of the target area and the possible placement positions. The method for proposing an arrangement of an electromagnetic sensor according to claim 1, wherein the electromagnetic sensor is an OPM (Optically Pumped Magnetometer) sensor, a fluxgate sensor, an MR (Magneto Resistance) sensor, an MI (Magneto Impedance) sensor, a coil-type sensor, or an NVC (Nitrogen-Vacancy Center in Diamond) sensor. The electromagnetic sensor includes:
2. The method for proposing an arrangement of electromagnetic sensors according to claim 1, wherein the sensors are magnetic sensors or electric potential sensors that detect magnetic fields generated by electrical activity in the brain, heart, spinal cord, peripheral nerves, or muscles. The method for proposing an arrangement of an electromagnetic sensor according to claim 1, wherein the step of outputting the proposed arrangement includes a step of presenting the proposed arrangement to a user. A method for estimating a position of a current source inside a subject using detection results of a plurality of electromagnetic sensors arranged on a biological surface of the subject, each of which detects an electromagnetic field generated by biological activity occurring in a target region inside the subject, comprising:
setting a calculation condition including a position of the target area and an arrangement of the plurality of electromagnetic sensors;
calculating priorities of all or a portion of the plurality of electromagnetic sensors based on the calculation conditions;
and estimating a position of the current source using detection results of the electromagnetic sensors selected based on the priority. 1. An apparatus for proposing an arrangement of a plurality of electromagnetic sensors to be placed on a biological surface around a target region inside a subject, using a forward model for estimating an electromagnetic field generated at the positions of the plurality of electromagnetic sensors due to biological activity inside the subject, the apparatus comprising:
an input unit to which calculation conditions including a position of the target area and possible positions of the plurality of electromagnetic sensors are input;
a calculation device that calculates a proposed arrangement of all or a part of the plurality of electromagnetic sensors for estimating a position of a current source inside the subject based on the calculation conditions;
and an output unit that outputs the proposed arrangement.

Images