JP2751408B2 - Magnetoencephalograph - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial applications]

The present invention relates to an apparatus that measures a magnetic field generated in a brain of a human or the like and estimates an active site of the brain.

[Prior art]

Conventionally, finding the position of the brain activity site by measuring the magnetic field generated in the human brain has been used in clinical medicine, such as estimating the location of epilepsy, researching spontaneous brain magnetism (especially alpha waves), and studying induced brain magnetism. (Ritta Mari and
Risto J. Ilmoniemi, “Cevebral Magnetic Fields” CRC
Critical Reviews in Biomedical Engineering Vol. 14,
No. 2 p93-126, 1986, L. Kaufman and J. Williamson, "Re
cent developments in neuromagnetism "in Third Inter
national Evoked Potentials Symposium, Stoneham: Butt
erworth, p100-113, 1987). Further, USP No. 4,736,751 proposes a system for identifying a brain active site by electroencephalogram measurement.

[Problems to be solved by the invention]

However, in the past, when measuring magnetoencephalography, a magnetic sensor with a small number of measurement points was used, so it was necessary to create a magnetoencephalogram by measuring the required range of the skull while changing the location of this magnetic sensor, It takes a long time to measure, and after creating a magnetoencephalogram, it simply creates an isomagnetic field map and displays it.Therefore, it is possible to grasp the entire magnetic field distribution or simply grasp the current dipole using the isomagnetic field map. However, there was a problem that accurate estimation of a large number of current dipoles was impossible. Also, USP No. 4,736.751 proposes a system suitable for electroencephalogram measurement.
Special measures are required to grasp the position of the sensor / MRI head image. An object of the present invention is to provide a magnetoencephalography measurement apparatus that does not require much time for magnetoencephalography and that can accurately determine the position of a brain activity site.

[Means for Solving the Problems]

In order to achieve the above object, in the magnetoencephalograph according to the present invention, a multi-channel SQUID (Superconducti
ng Quantum Interference Device) A means for simultaneously measuring brain magnetic fields at multiple points using a sensor, and an MRI means for taking MRI images on a number of slice planes parallel to a reference plane passing through feature points on the head. And a magnetic field formed by a small coil through which a current pulse is provided at the characteristic position of the head, measured by the multi-channel SQUID sensor, and a reference passing through the characteristic point of the head obtained by the MRI means. Means for determining the positional relationship between the multi-channel SQUID sensor and the MRI image using image data on a number of slice planes parallel to the plane; means for creating a head approximation model from the MRI image; and The position, size, and orientation of a plurality of current dipoles are assumed in, and the magnetic field distribution created by the current dipole group at the measurement point of the brain magnetic field and the multi-point brain magnetic field measured simultaneously Means the difference between the magnetic field distribution obtained from the chromatography data seek current dipole groups as to minimize the said current dipole group and means for displaying on the image associated with the MRI images is provided.

[Action]

By using a multi-channel SQUID sensor,
Brain magnetic fields can be measured simultaneously at multiple points. On the other hand, an MRI image of the head is taken by MRI means. At this time, image data is obtained on a number of slice planes parallel to a reference plane passing through the feature points of the head. Then, a multi-channel SQUID sensor measures a magnetic field formed by a small coil provided with a current pulse, provided at a characteristic position of the head, and the magnetic field measurement data,
The positional relationship between the multi-channel SQUID sensor and the MRI image is obtained by using the image data obtained on the multiple slice planes parallel to the reference plane passing through the feature points of the head and obtained by the MRI means. Thus, the positional relationship between the magnetoencephalographic measurement data and the measurement point MRI image can be determined. A head approximation model was created from the MRI image.In this model, the position, size, and direction of a plurality of current dipoles were assumed, and the current dipole group and the simultaneous By obtaining a group of current dipoles such that the difference from the magnetic field distribution obtained from the measured multi-point brain magnetic field data is minimized, each position of the current dipole,
You can see the size and direction. Each position of the current dipole group obtained in this way corresponds to the MRI image, and thus can be displayed on an image related to the MRI image (that is, the MRI image itself or an image created therefrom). Thereby, the brain activity site is displayed on the MRI image as a current dipole group, and the position can be known precisely.
As described above, the magnetic field formed by the small coil provided at the characteristic position of the head is measured by the multi-channel SQUID sensor, and the MRI means is used to measure the number of slice planes parallel to the reference plane passing through the characteristic point of the head. MRI image of
With these, the position information of the magnetoencephalogram data and the position information of the MRI data are matched, so that the positional relationship is accurate, and the positional accuracy of the resulting brain activity striking position is increased. The operation of collating is also easy.

【Example】

Next, an embodiment of the present invention will be described with reference to the drawings. As shown in FIG. 1, a subject 10 undergoes an examination in an examination room (a shield room or a room with a small external magnetic field) 8. All the structures and the like used in the inspection room 8 are made of a non-magnetic material. A multi-channel SQUID sensor 11 is set on the head of the subject 10. A display device 12 such as a TV monitor device is arranged outside the examination room 8, and a visual stimulus is given to the subject 10 by observing the image via a mirror 12a. In addition, earphone 13
Is placed outside the examination room 8 and its sound is guided to the ear of the subject 10 through the silicon tube 13a or the like, thereby giving an auditory stimulus. Further, a somatic sensory stimulus is given using an electric stimulator (not shown), or a motion detector 14 is used to perform a recognition operation. These are controlled by the experiment control device 15, and the measurement system 9 is formed. The data collection device 16 collects multipoint brain magnetic field measurement data from the multi-channel SQUID sensor 11. Subject
The head of 10 is imaged by the MRI apparatus 17. The multipoint measurement data of the brain magnetic field obtained by the data collection device 16 and the MRI image data obtained by the MRI device 17 are
Sent to In the computer 18, the magnetic field measurement data is not subjected to special processing in the case of epileptic magnetoencephalography, is subjected to FFT (Fast Fourier Transform) processing in the case of spontaneous electroencephalogram such as α-wave, and the S / N ratio is Undergoes an addition (integration) process to improve. In each of these cases, digital filtering may be further performed to improve the S / N ratio. An approximate head model is created from the head MRI image. Further, the relative positional relationship between the SQUID sensor and the head is determined, and the positional relationship between the magnetoencephalographic measurement point and the head MRI image is determined. After the positional relationship between the measurement point and the MRI image is grasped in this way, the approximate head model is selected, and the coordinate system is selected. Thus, the coordinate system is determined, and the position, size, and direction of each current dipole group are obtained using the magnetoencephalographic data subjected to the above processing. This is equivalent to solving the reverse problem.Here, the positions, sizes, and directions of a plurality of current dipoles are assumed on the coordinates described above, and the magnetic field distribution created by the current dipole group at the magnetic field measurement point is calculated. Each position, size, and direction of the current dipole group that minimizes the least square error with the actually measured magnetic field distribution is obtained. The current dipoles determined in this way are displayed on the MRI image or a three-dimensional image created therefrom with a mark such as an arrow on the display device 19, or displayed as coordinates. The multi-channel SQUID sensor 11 is cooled by being immersed in liquid helium filled in a container usually called a dewar. Here, as shown in FIG. 2, three dewars 2 are used at the same time, and the bottom of the tip of the dewar 2 surrounds the subject's head. Each of these three dewars 2 is attached to the tip of an arm 22 which extends and contracts from the base 21 so as to extend freely.
It is held by a belt 24 via 23. The dewar 2 has a substantially circular shape, and the belt 24 can be rotated about its central axis by loosening a fixing member (not shown). Also, this belt 24 is
The arm 22 is rotatably attached to the tip of the arm 22. Further, the base 21 is movable along a rail (not shown) provided on the ceiling and is rotatable about a vertical axis. With such a suspending mechanism, each dewar 2 can be freely positioned in the horizontal and vertical directions, and can be freely set in a horizontal plane.
The tilt angle with respect to the vertical axis can be set freely,
It can also rotate around its own central axis. In addition, the bottom surfaces of the dewars 2 located on both sides are inclined at about 45 ° from the central axis, as shown in FIGS. The bottom surface of the central dewar 2 is substantially perpendicular to its central axis. Each of these bottom surfaces is a spherical concave surface similar to the skull. The interior of the dewar 2 is configured as shown in FIG.
The dewar 2 includes a liquid helium container 32 in which the liquid helium 41 is contained, and an envelope 31 that covers the outside thereof. A vacuum 33 is formed in the envelope 31, and the liquid helium container 32
A super-insulation layer 35 as a heat insulating material and a steam-cooled metal strip 34 are arranged so as to surround.
A number of detection coils 42 are arranged at the bottom of the liquid helium container 32. The detection coils 42 are respectively wound around quartz bobbins. Since the dewar 2 has a bottom surface inclined at 45 ° and a spherical concave surface, a large number of detection coils 42 are arranged along the 45 ° concave surface. In the case of the central dewar 2, the bottom surface is a substantially right-angled spherical concave surface, and a large number of detection coils are arranged along the bottom surface. In any case, when the detection coils 42 are viewed from the front, they are arranged at equal intervals as shown in FIG. 4, and are directed to the center of curvature of the spherical concave surface. Here, 19 detection coils 42 are arranged per dewar as shown in FIG. 4, and they are connected to 19 DC SQUID units 43, respectively. These detection coil 42 and DC
The SQUID unit 43 is supported by a support cylinder 44. A heat radiation shield 45 is attached to the upper part of the support cylinder 44. The magnetic flux detected by the detection coil 42 is transmitted to the DC SQUID unit 43 by a magnetic flux transformer, and the FLL (Flux Lo
The magnetic flux is converted into a voltage by the operation of the ck Loop), and the voltage signal is output to the outside via a preamplifier unit 46 attached outside the dewar 2. Since each dewar 2 has such a concave bottom surface and the position and direction of the dewar 2 can be freely determined as described above, the detection coil
The position 42, that is, the magnetic field measurement point, can be on a spherical surface approximating the skull. The 57 (= 19 × 3) measurement signals are transmitted to the data collection device 16
Is sent to the computer 18. That is, as shown in FIG. 5, the data is sent to the CPU bus 52, stored in the CPU memory (not shown), and
The data is sent to the hard disk driver 54 via the storage 53 and stored. Thereafter, processing by the CPU 51 is performed, and the obtained image is stored in the image memory 58, and then sent to the color image monitor device 59, where the image is displayed.
In addition, a character monitor device 56 and a keyboard device 57 are connected to the CPU bus 52 via the console controller 55, so that data processing procedures and the like can be input. The data collection device 16 is configured as shown in FIG. SQUI
The signal (analog) sent from the FLL of the D sensor 11 is first captured via the amplifier 61, and then input to the A / D converter 65 via the filter 62, the sample-and-hold circuit 63 and the multiplexer 64, where it is digitally converted in real time. After being converted to a signal,
The signal is output through a FIFO (First In First Out) circuit 66. On the other hand, the positional relationship between the magnetic field measurement point and the MRI image is obtained as follows. Several small coils as shown in Fig. 7
71 is attached to the characteristic position of the skull of the subject 10. As the characteristic positions, three points out of the root of the nose, the notch in front of the auricle, and the outer occipital ridge are desirable. A known current pulse is sequentially passed to each of the small coils 71 attached in this manner, and a plurality of
It is measured by the sensor 11. By storing in advance in the computer 18 the magnetic field measured by the small coil 71 placed on known coordinates, the small coil attached to the skull
71 positions can be obtained. This allows small coils
The positional relationship on the magnetic field measurement coordinates of the skull feature point to which 71 is attached can be known. Instead of using the small coil 71 as described above, 3
The positional relationship between the measurement point of the SQUID sensor 11 and the head can also be obtained using a dimensional digitizer. The three-dimensional positional relationship between the head and the SQUID sensor 11 is obtained by attaching a quadrature coil to the head side, generating a quadrature magnetic field by the quadrature coil, and detecting the quadrature coil provided on the SQUID sensor 11 side. Is to be detected. On the other hand, in the MRI apparatus 17, as shown in FIG. 8, a plane passing through the root of the nose, the notch in front of the auricle, and the occipital ridge is defined as an xy plane, and a large number of slice planes parallel to the xy plane are defined. (Xy
(Arranged in the z direction perpendicular to the plane). Since this image data is sent to the computer 18, the positional relationship between the MRI image and the magnetic field measurement data can be matched. In computer 18 this three-dimensional
From the MRI image data, boundaries of the scalp, skull, and brain are identified, and an approximate model is created by using these boundaries. This approximate model is shown in, for example, FIG.
This is as shown in FIG. 11 and FIG. FIG. 9 represents the cerebral cortex as a homogeneous sphere. FIG. 10 and FIG. 11 show the scalp epidermis, scalp epidermis-skull boundary, skull-cerebral cortex boundary,
Expressed by triangular elements. The triangular element representation is the most precise model. Assuming the positions, sizes, and directions of a plurality of current dipoles in these models, a so-called backward problem is solved. The magnetic field Bir created by the current dipole at the measurement point is a spherical model (Fig. 9)
In the case of, an analytical expression is already required (B.NEIL CUF
FIN et al. “Magnetic Fields of a Dipole in Special
Volume Conductor Shapes "IEEE Trans.Biomed.Eng., Vo
l. BME-24, No. 4, pp. 372-381, 1977). In the case of the multilayer sphere model (Fig. 10) and the precision model (Fig. 11), the model is represented by triangular elements,
Boundary element method (CA Brevia "Introduction to Boundary Element Method" Baifukan), boundary integral equation method (AC
L. Barnard et al. “The application of electromagnet
ic theory of electrocardiology "Biophysical Journa
1, Vol. 7, p443-491, 1967). Then, a least square error E with respect to the measured value Bie is obtained as follows. Here, N is a measurement point. The positions, the sizes, and the directions of the plurality of current dipoles that minimize the least square error E are determined. When the position, size, and direction of each current dipole are obtained in this way, they are superimposed on a three-dimensional MRI image or on a three-dimensional image obtained by transforming the three-dimensional MRI image with marks such as arrows. . Thus, for example, an arrow representing a current dipole is displayed on a three-dimensional image of the head as shown in FIG. 12, or as shown in FIG. 13A, B, C, a three-dimensional sectional view of coronal, sagittal, and transverse. An arrow is displayed on. Further, the distance and coordinates of the current dipole from any point on the epidermis of the head may be displayed. In addition, not only the current dipole corresponding to the brain activity at that time is displayed from the magnetoencephalogram measurement data at a certain time, but also the brain magnet is continuously measured for a long time, and then the current dipole is obtained for a certain period of time. It is also possible to know the temporal change of the brain activity and perform a dynamic grasp of the brain activity, for example, by displaying the information and changing the color of the current dipole every time.

【The invention's effect】

According to the magnetoencephalograph of the present invention, the multi-channel
Since the brain magnetism measurement is performed with the SQUID sensor, the inspection time can be reduced. In addition, since the collation between the magnetoencephalographic measurement data and the MRI image data by the multi-channel SQUID sensor is easy and accurate, the precise position of the brain activity site can be obtained. Therefore, for example, in the case of epilepsy, the positional correspondence between the abnormal biofunction on the MRI image and the current dipole generated by the epilepsy becomes clear.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an entire system according to an embodiment of the present invention, FIG. 2 is a front view showing an arrangement of a dewar in which a SQUID sensor is housed, FIG. The figure shows the positional relationship of the detection coils.
FIG. 6 is a block diagram showing a hardware configuration of a computer, FIG. 6 is a block diagram showing a hardware configuration of a data collection device, FIG. 7 is a diagram showing an attachment position of a small coil to a skull of a subject, and FIG. FIG. 9 shows a skull sphere model, FIG. 10 shows a multi-layer skull model of the skull, and FIGS. 11A, 11B and 11C show precision simulated homogeneous models of the skull. FIG. 12 is a diagram showing a three-dimensional image of a head cut, and FIGS. 13A, 13B, and 13C are diagrams showing cross-sectional images of the head in three directions. 8 ... laboratory, 9 ... measurement system, 10 ... subject, 11 ... SQ
UID sensor, 12, 19 …… Display device, 12a …… Mirror, 13…
… Earphone, 13a …… Silicon tube, 14 …… Motion detector, 15 …… Experiment control device, 16 …… Data collection device, 17
…… MRI device, 18 …… Computer, 2 …… Dewar, 2
1 ... base, 22 ... arm, 23 ... shaft, 24 ... belt, 31 ... envelope, 32 ... liquid helium, 33 ... vacuum
34 ... Steam cooled metal strip, 35 ... Super insulation layer, 41 ... Liquid helium, 42 ... Detection coil, 43 ... DC SQUID unit, 44 ... Support cylinder, 45 ...
Heat radiation shield, 46 …… Preamplifier unit, 51 …… CP
U, 52: CPU bus, 53: Disk interface,
54: Hard disk drive device, 55: Console controller, 56: Character monitor device, 57:
Keyboard device, 58 Image memory, 59 Color image monitor device, 61 Amplifier, 62 Filter, 63 Sample hold circuit, 64 Multiplexer, 65 A / D converter, 66 ... FIFO circuit, 71 ... Small coil.

Claims (1)

(57) [Claims] A multi-channel SQUID sensor for simultaneously measuring brain magnetic fields at multiple points; an MRI means for capturing MRI images on a number of slice planes parallel to a reference plane passing through feature points of the head; Data obtained by measuring the magnetic field formed by the small coil through which the current pulse is provided at the characteristic position of the section with the multi-channel SQUID sensor,
Means for obtaining a positional relationship between the multi-channel SQUID sensor and the MRI image by using image data on a number of slice planes parallel to a reference plane passing through the feature points of the head obtained by the RI means; and Means for creating a head approximation model from an image, positions of a plurality of current dipoles in the model,
Assuming the size and direction respectively, the difference between the magnetic field distribution created by the current dipole group at the measurement point of the brain magnetic field and the magnetic field distribution obtained from the simultaneous measurement of multiple points of the brain magnetic field data is minimized. A magnetoencephalography apparatus comprising: means for obtaining a current dipole group; and means for displaying the current dipole group on an image related to the MRI image.