JP4899121B2 - Current unit estimation method and apparatus - Google Patents

Description

  The present invention relates to a current element estimation method and apparatus, and more particularly, to a current element estimation method and apparatus that can reduce the computation time.

Conventionally, various methods have been proposed for estimating the distribution of current elements (equivalent current dipoles) assumed to exist in the brain from magnetic field data measured by a brain magnetism detection device (for example, Non-Patent Document 1, Non-Patent Document 1). Reference 2).
Also, a current element localization processing method for narrowing down a plurality of current elements that generate a measurement magnetic field is known (see Patent Document 1).
Also, a biomagnetism measuring device that measures the magnetic field of the spinal cord is known (see Patent Document 2).
Co-authored by Hiroshi Hara and Shinya Kurishiro, “Neuromagnetic Science, SQUID Measurement and Medical Applications”, Ohm Publishing House, January 25, 1997, p. 101-120 The Institute of Electronics, Information and Communication Engineers, Takeda Tsunehiro, “Electronic Information and Communication Lecture Series D-24, Brain Engineering”, Corona, April 11, 2003, p. 128-140 JP 2005-058558 A JP 2005-337862 A JP 2005-058558 A

In the conventional method of estimating the distribution of current segments, the number of current segments is unnaturally limited due to the convenience of calculation, or a local solution (which is unlikely to be practical even if the number of current segments is increased) For example, there is a problem that a satisfactory estimation result cannot be obtained, such as falling into a solution in which a current segment exists in a place where it is unlikely to exist.
Further, in the conventional current element localization processing method, it is necessary to narrow down six elements of the position x, y, z, the current magnitude j, and the current directions θ and ψ for one current element. In addition, if there are N current pieces, it is necessary to adjust 6 × N elements, and there is a problem that it takes a long calculation time.
Therefore, an object of the present invention is not to limit the number of current elements unnaturally, nor to fall into a local solution unlikely in practice, and to adjust current elements by adjusting a small number of elements. It is an object to provide a current segment estimation method and apparatus capable of estimating the current.

In a first aspect, the present invention provides a curve generating step for generating a smooth curve in a three-dimensional space, a current element generating step for generating N current elements distributed on the curve, and a current element. And a search step for changing the element of the current element so that the error between the calculated magnetic field calculated from the distribution of the pieces and the measured magnetic field is sufficiently small, and the error is less than the allowable value by repeating part or all of the steps. A current element estimation method is provided, characterized in that a current element is obtained.
For example, when the magnetic field of the brain is measured and the current element is estimated based on the measurement data, the current element is distributed in the brain, that is, in a three-dimensional volume, so the position and the magnitude and direction of the current are limited. It ’s difficult. However, for example, when a magnetic field emitted from the spinal cord is measured and the current element is estimated based on the measured data, the current element is present on the spinal cord, ie, a generally S-shaped curve. You can easily limit the direction.
Therefore, in the current unit estimation method according to the first aspect, first, a curve in a three-dimensional space is generated. For example, a Bezier curve or a spline curve can be used for the expression of the curve. If this curve is, for example, S-shaped, it can be approximated by a cubic curve, so the curve can be defined by four points in space. Next, N current pieces distributed on the curve are generated, and the position on the curve can be defined by one parameter k. Further, since the direction of the current is the tangential direction of the curve, it is automatically determined when the position of the current element is determined, and does not need to be adjusted independently. Therefore, it is sufficient to narrow down the two elements of the position k and the current magnitude j for one current element, and if the number of current elements is N, adjustment of 2 × N elements is sufficient, and the calculation time is shortened. I can do it. However, since it is necessary to adjust, for example, four points that define the curve, it is not reduced to 1/3 compared to the conventional 6 × N. However, since the adjustment of the points that define the curve is not multiplied by the number of current segments (N times), the overall calculation time can be reliably shortened as compared with the conventional method.
In the above configuration, “repeating part or all of the steps” means that there are both cases where the curve generation step does not need to be repeated and where the curve generation step needs to be repeated.

In a second aspect, the present invention provides the current unit estimation method according to the first aspect, wherein only the magnitude of the current is changed in the searching step. .
In the current element estimation method according to the second aspect, for example, N current elements that are evenly distributed on the curve are generated, and the position is not moved, but only the magnitude of the current is changed and the search is performed. To do. Thereby, the calculation time can be further shortened.

In a third aspect, the present invention provides the current unit estimation method according to the first or second aspect, wherein the measurement magnetic field is a magnetic field emitted from a spinal cord. provide.
In the current element estimation method according to the third aspect, the current element is estimated based on the magnetic field data generated from the spinal cord. Since the spinal cord is generally an S-shaped curve, the present invention is preferably used. Applicable.

In a fourth aspect, the present invention provides a curve generating means for generating a smooth curve in a three-dimensional space, a current element generating means for generating N current elements distributed on the curve, and a current element. Search means for changing the element of the current element so that the error between the calculated magnetic field calculated from the distribution of the pieces and the measurement magnetic field is sufficiently small, and the error is allowed by operating part or all of the means repeatedly. Provided is a current unit estimation apparatus that obtains a current unit that is equal to or less than a value.
In the current unit estimation apparatus according to the fourth aspect, the current unit estimation method according to the first aspect can be suitably implemented.

In a fifth aspect, the present invention provides the current unit estimation apparatus according to the fifth aspect, wherein the search means changes only the magnitude of the current.
In the current unit estimation apparatus according to the fifth aspect, the current unit estimation method according to the second aspect can be suitably implemented.

In a sixth aspect, the present invention provides the current unit estimation apparatus according to the fourth or fifth aspect, wherein the measurement magnetic field is a magnetic field emitted from a spinal cord. provide.
In the current unit estimation apparatus according to the sixth aspect, the current unit estimation method according to the third aspect can be suitably implemented.

  According to the current element estimation method and apparatus of the present invention, the number of current elements is not unnaturally limited, and does not fall into a local solution that is unlikely to be practical, and a small number of elements. It is possible to estimate the current segment by adjusting. That is, the calculation time can be shortened.

  Hereinafter, the present invention will be described in more detail with reference to embodiments shown in the drawings. Note that the present invention is not limited thereby.

FIG. 1 is a configuration diagram illustrating a current unit estimation apparatus 100 according to Embodiment 1 of the present invention.
The current element estimation apparatus 100 includes a data collection unit 1 that collects biomagnetic field data generated from the spinal cord h of the subject H via the biomagnetism measurement apparatus A, and a distribution of current elements from the collected magnetic field data. Current segment estimation processing unit 2 and display unit 3 that visualizes and displays the estimated current segment.

FIG. 2 is a flowchart showing a procedure of current unit estimation processing by the current unit estimation device 100.
In step S1, the data collection unit 1 obtains magnetic field data represented by a magnetic field vector in a three-dimensional space, for example, as indicated by an arrow in FIG.

In step S2, the current element estimation processing unit 2 specifies a start point Q0, a control point Q1,... Defining a Bezier curve in a preset search space (a space as small as possible where a current element is present). And the end point QR is set at random. For example, as shown in FIG. 4, a start point Q0, control points Q1, Q2 and an end point Q3 are set.
In step S3, the current element estimation processing unit 2 generates an R-order Bezier curve from the start point Q0, the control point Q1,... And the end point QR. For example, as shown in FIG. 5, a cubic Bezier curve B is generated from the start point Q0, the control points Q1, Q2 and the end point Q3.

In step S4, N current pieces J1 to JN are generated randomly distributed on the R-order Bezier curve. For example, as shown in FIG. 6, eight current pieces J1 to J8 are generated.
In the cubic Bezier curve when R = 3, the positions xn, yn, zn of the current element Jn are expressed by the following equations.
^{xn = (1-kn) 3} x0 + 3kn (1-kn) 2 x1 + 3kn 2 (1-kn) x2 + kn 3 x3
^{yn = (1-kn) 3} y0 + 3kn (1-kn) 2 y1 + 3kn 2 (1-kn) y2 + kn 3 y3
^{zn = (1-kn) 3} z0 + 3kn (1-kn) 2 z1 + 3kn 2 (1-kn) z2 + kn 3 z3
Here, kn is a parameter representing the position of the current element Jn on the R-order Bezier curve, and 0 ≦ kn ≦ 1.
Further, the position coordinates of the point Qr are assumed to be xr, yr, zr. For example, the position coordinates of the starting point Q0 are x0, y0, z0.
The cubic Bezier curve is described in, for example, Keizo Sakurai et al., “Introduction to Graphic Processing-Approach to CAD / CG”, Asakura Shoten, 2002, pp. 59-61, “2.4.2 Bezier Curve”. .

The current direction of the current element Jn is the direction of the tangent vector, and is expressed by the following equation.
xn '= - 3 (1- kn) 2 x0 + 3 (3kn-1) (1-kn) x1 + 3kn (2-3kn) x2 + 3kn 2 x3
yn '= - 3 (1- kn) 2 y0 + 3 (3kn-1) (1-kn) y1 + 3kn (2-3kn) y2 + 3kn 2 y3
zn '= - 3 (1- kn) 2 z0 + 3 (3kn-1) (1-kn) z1 + 3kn (2-3kn) z2 + 3kn 2 z3

  The current magnitude jn of the current element Jn is determined randomly within a possible range.

In step S5, as shown in FIG. 7, the current element estimation processing unit 2 calculates magnetic vectors (arrows in FIG. 7) generated by the N current elements J1 to JN from, for example, the biosavart equation.
In step S6, the current element estimation processing unit 2 calculates an error between the magnetic field data (FIG. 3) and the calculated magnetic field (FIG. 7) using a known evaluation function. If the error is not sufficiently small, the process proceeds to step S7. If the error is sufficiently small, the process proceeds to step S11.

  In step S7, the current element estimation processing unit 2 determines whether or not the error is likely to be small (for example, if the number of times that the process has passed through step S6 is less than 10,000, the error is likely to be small). If it is 10,000 times or more, it is determined that there is no possibility that the error will be reduced.) If there is a possibility, the process proceeds to step S8.

In step S8, the current element estimation processing unit 2 randomly selects one current element Jn.
In step S9, the current element estimation processing unit 2 selects a parameter to be changed. That is, the parameter kn indicating the position of the current element Jn and the current magnitude jn are selected, or only the current magnitude jn is selected.
In step S10, the current element estimation processing unit 2 slightly changes the selected parameter of the selected current element Jn in a direction to reduce the error using a known optimization method. Then, the process returns to step S5.

  By repeating the above steps S2 to S10 a sufficient number of times, the distribution of current segments Jn with a sufficiently small error is determined.

  In step S11, as shown in FIG. 8, the current element estimation processing unit 2 visualizes and displays the magnetic field data and the distribution of the N current elements Jn.

  According to the current unit estimation apparatus 100 of the first embodiment, the number N of current units Jn is not unnaturally limited, and does not fall into a local solution that is unlikely to be practical, and a small number The current element Jn can be estimated by adjusting the elements. That is, the calculation time can be shortened.

  In step S2 of FIG. 2, the start point Q0, the control point Q1,... And the end point QR are set randomly at the first time, but the immediately preceding start point Q0, control point Q1,. It is also possible to change the curve gradually.

  In step S2 of FIG. 2, the start point Q0, the control point Q1,... And the end point QR are set on a straight line that passes through the center of the search space for the first time. The curve may be gradually changed by slightly changing the points Q1,... And the end point QR.

  The shape and position of the spinal cord are detected in advance using an MRI image, an X-ray image, etc., and in step S2 in FIG. 2, the first time is the start point Q0, the control point Q1,. QR may be set, and in the second and subsequent steps S2, the curve may be gradually changed by slightly changing the immediately preceding start point Q0, control point Q1,.

  In step S4 in FIG. 2, N current pieces J1 to JN may be generated by being evenly distributed on the curve.

In step S4 of FIG. 2, the magnitude jn of the current element Jn may be determined by the lead field method using magnetic field signal information from a plurality of magnetic sensors.
Regarding the lead field method, for example, Matti Hamalainen et al "Magnetoencephalography-theory, instrumentation, and applications to noninvasive studies of the working human brain" Review of Modern Physics Vol.65 No.2, April 1993, p.413-498 It is described in.

Instead of using a cubic Bezier curve, a cubic spline curve, a quadratic Bezier curve, or the like may be used.
These curves are described in, for example, Keizo Sakurai et al., “Introduction to Graphic Processing-Approach to CAD / CG”, Asakura Shoten, 2002, “2.4 Free Curves”.

  For example, it can be used to estimate the distribution of current segments assumed to exist in the spinal cord from data obtained by measuring a magnetic field generated from the spinal cord with a biomagnetic measuring device.

1 is a configuration diagram illustrating a current element estimation device according to Embodiment 1. FIG. It is a flowchart which shows the procedure of an electric current unit estimation process. It is a conceptual diagram of magnetic field data. It is a conceptual diagram of the point which prescribes | regulates a curve. It is a conceptual diagram which shows the curve which generate | occur | produced. It is a conceptual diagram which shows the generated electric current piece. It is a conceptual diagram which shows a calculation magnetic field. It is a conceptual diagram which shows distribution of the visualized magnetic field data and an electric current piece.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Data collection part 2 Current unit estimation process part 3 Display part 100 Current unit estimation apparatus A Biomagnetic measurement apparatus H Subject h Spinal cord

Claims (6)

A curve generating step for generating a smooth curve in a three-dimensional space; a current element generating step for generating N current elements distributed on the curve and having a tangential direction of the curve as a current direction; And a search step for changing the element of the current element so that the error between the calculated magnetic field and the measurement magnetic field is sufficiently small. A current unit estimation method characterized by obtaining a current unit that repeats part or all of the error and makes the error equal to or less than an allowable value.   The current unit estimation method according to claim 1, wherein, in the searching step, only the magnitude of the current is changed.   3. The current unit estimation method according to claim 1, wherein the measurement magnetic field is a magnetic field emitted from a spinal cord. Curve generating means for generating a smooth curve in a three-dimensional space; current element generating means for generating N current elements distributed on the curve and having a tangential direction of the curve as a current direction; Search means for changing the element of the current element so that the error between the calculated magnetic field calculated from the current element distribution in which the position of the piece is defined by one parameter and the measurement magnetic field is sufficiently small. A current unit estimation apparatus characterized by obtaining a current unit that operates part or all repeatedly to make the error equal to or less than an allowable value.   5. The current unit estimation apparatus according to claim 4, wherein the search unit changes only the magnitude of the current.   6. The current unit estimation apparatus according to claim 4 or 5, wherein the measurement magnetic field is a magnetic field emitted from a spinal cord.

Images