FI121899B - Method, device and computer software product for magnetic image making of objects - Google Patents

Description

Method, hardware and computer program product for magnetic imaging of objects

The invention relates to magnetic imaging of objects, for example, by magnetic resonance imaging (MRI). In particular, the invention relates to encoding MRI signals so as to obtain an image of the subject, i.e. a method of imaging MRI according to the preamble of claim 1. The invention further relates to an imaging device and a computer program product for performing such a method on a MRI device.

Magnetic resonance imaging (MRI) is a non-destructive method for studying the internal structure of matter. The description is based on the magnetization of the sample and the observation of the precursor signals produced by the magnetized sample. Measured signals can be used to produce an image of an object.

15

Magnetic resonance imaging techniques can be roughly divided into two categories: imaging using the same main and metering magnetic field, and pre-polarized imaging. In the first case, the same main magnetic field is used as both the polarization field of the sample and the precursor field of the signals (measurement field). This field is very homogeneous in the sample region 20 and causes the sample to polarize magnetically in the direction of the field, i.e. net magnetization to the sample. In order to measure net magnetization, it must be turned away from the polarization axis. This is done with a radio frequency RF pulse which generates a variable magnetic field component perpendicular to the field lines of the main magnetic field (the so-called Larmor frequency of the atomic nucleus of the substance, i.e. the precession frequency is proportional to? 25 magnetic field amplitude). When the RF pulse is switched off, the net magnetization begins around the directions of the main magnetic field to return the precessors toward the oo polarization direction. This so-called during relaxation, the preconception of net magnetization can be measured x by one or more receiving coils or sensors outside the sample.

CL

Typically, the so-called. high-field imaging, where the magnitude of the main magnetic field can be as high as several £ 5 to 30 tesls, falls into this category. Also the so-called. for low-field imaging, the 00 o term RF pulse is used similarly to high-field imaging, even though the precession frequencies C11 are not in the radio frequency range.

2

In the second case, i.e., in the pre-polarized MRI, the sample is initially polarized in a polarization field, after which the polarization field is switched off and a measuring field lower than the polarization field is typically switched on. The preconception of the net magnetization around the measurement field (typically including various gradient fields) is measured outside the 5 samples with one or more receiving coils (or sensors). Pre-polarized imaging can be performed completely without RF pulses or by utilizing RF pulses to reverse magnetization from the polarization direction.

For low-field pre-polarized MRI, superconducting quantum interference device (SQUID) sensors or other sensors such as mixed sensors or atomic magnetometers can be used. In pre-polarized low-field MK, the sample is initially polarized in a magnetic field, for example, in the order of tens of milliseconds. The polarization field is then switched off, and a lower (for example, some microtesl) size field is turned on at least 15 partially orthogonal to the direction of the polarization field. Again, the precedence of net magnetization around the measuring field is measured from the outside of the sample by one or more receiving coils (or sensors). On the other hand, low-field imaging can also be performed without pre-polarization using suitable sensors.

20 Since the rate of relaxation depends on the physical properties of the substance, and is therefore slightly different from one substance to another, information on the internal structure of the substance is obtained if the origin of the measured MK signals can be located. So this is the solution to the inverse problem. Conventionally, Fourier coding has been used to locate MK signals. In the method, linear gradient fields of different magnitudes are added to the measurement field, whereby the pre-o 25 processor frequencies become location dependent (frequency coding). On the other hand, by briefly exposing the pre-cessing magnetization to a gradient field, the precession phase becomes location dependent (phase coding). Particularly in high field MK, the way is also to deflect magnetization by RF pulse when a gradient field is applied over the sample, whereby only

CL

The portion of the sample in the field corresponding to the RF frequency is magnetized (slice coding). The methods mentioned above for r5 30 are very popular and the imaging can be done with a fast Fourier transform. At present, the measurement rate of Fourier coding is limited by the available-

(M

gradients and safety factors. Low magnetic fields give higher fields more opportunities for measurements and design of imaging sequences.

3

The magnitude of the object's magnetization, and hence the magnitude of the precession signal to be measured, depends on the magnitude of the magnetic field used, and the formation of an accurate image takes considerable time in the shallower fields, especially if the noise level is significant. As a result, in magnetic resonance imaging, especially in low-field MK, the measurement times 5 are long and there is a need in the art for methods to accelerate imaging.

Recently, the use of multiple receiving coils in magnetic resonance imaging has become more common, since the information provided by the sensitivity profiles of the coils makes it possible to shorten measurement times. These methods are called parallel MRIs. On the other hand, several coils 10 can also be used without speeding up the imaging, for example to reduce the noise level.

General information on MRI and imaging is presented e.g. in Z.-P. Liang and P. C. Lauterbur. Principles of Magnetic Resonance Imaging: A Signal Process-15 ing Perspective. IEEE Press Series in Biomedical Engineering. IEEE Press, 2000.

The object of the invention is to provide a new method and apparatus for imaging in magnetic resonance imaging, in particular in pre-polarized magnetic resonance imaging. In particular, it is an object of the invention to provide a method which provides more information about a subject in a given time and, on the other hand, enables the measurement times to be shortened. It is also an object of the invention to provide, or at least reduce, the need to use gradient fields during signal measurement.

The objects of the invention are achieved by a method according to the independent claim 1.

The invention is based on the idea that phase and / or Q_ amplitude differences of magnetization are produced in the sample already during the polarization of the sample. Thus, unlike conventional r5 30, polarization is not attempted to be performed over the entire sample in the same way (homogeneously) o at each polarization, but deliberately and controlledly produces different polarization distributions on sample C 1, most of which are typically inhomogeneous. Namely, according to the invention, it has been found that by repeating the measurement using different polarization distributions, 4 new samples are obtained which are suitable for use in imaging. In other words, when the polarization field is changed, the initial polarization of the object is different at different iterations of the measurement. When the variations in the polarization of a sample between different times, i.e. the position dependence of net magnetization, are known, an image of the object can be formed.

5

The invention thus provides an imaging method for magnetic imaging of a subject, the method comprising (a) generating a magnetic polarization to the target, (b) measuring the magnetization of the target by at least one receiving coil or sensor, and (e) generating at least one dimensional image of the object. According to the invention, - steps (a) and (b) are repeated several times using a plurality of polarizing magnetic fields having substantially different position-dependent profiles in the subject, and a 15 image of the object is formed based on said measurements of a polarized sample with multiple polarizing magnetic fields.

The apparatus for magnetic imaging of an object according to the invention comprises: - means for producing measurable magnetic polarization on the object, - means for measuring at least one sensor generated polarization outside the object, and - means for generating measurable polarization on the object;

5 25

CM

The means for producing magnetization with different polarization profiles may comprise, for example, a plurality of coils of various shapes (which generate different fields of magnetic fields), movable or deformable coils, movable permanent magnets, or various

CL

via magnetic fields (such as metal plates, etc.). The movement of the coils can be effected by translation and / or rotation. On the other hand, different polarization profiles can also be achieved by using at least one non-homogeneous magnetic field and

CM

la in this field by translation and / or rotation. Also in this case, the shape of the magnetic field in the target area varies as desired. Combinations of the methods described above are also possible. If the object is moved, the measuring sensors can also be moved with the object, but this is not necessary.

The magnetization measurement is typically performed in a non-directional field of measurement, i.e., it has a perpendicular component relative to the direction of the polarization fields. The fields may also be substantially perpendicular to one another. On the other hand, parallel fields can also be used if the magnetization direction is changed after polarization by RF pulses. The measurement field is typically substantially homogeneous in the imaging region, especially if polarization coding is combined with Fourier coding.

10

In image acquisition, information about the shape of the polarizing magnetic fields used in the measurement is used to encode, for example, one dimension of the object, i.e., to calculate voxel intensities in this dimension. Clip, phase, or frequency coding may be used to encode at least one other dimension of the object. It is also possible to encode any of the 15 dimensions both using different polarization fields as well as phase or frequency coding. Sensitivity profiles of different sensors are also used in polarization coding for coding.

According to one embodiment, polarizing magnetic fields are used that are at least substantially non-linearly variable in the region of the object 20. In this respect, too, the method differs substantially from traditional phase coding, which uses a linear gradient field arranged over a measuring field.

In the pre-polarized low-field imaging, each repetition, the polarization field is turned on and off before the measurement begins in a measuring field that is typically homogeneous in the sample region. In pre-polarized low-field imaging, polar-

The K i field strength is typically less than 100 mT and the measurement field is typically less than 1 mT.

According to one embodiment, the present apparatus comprises means for moving the object cc Q_ relative to the polarization and signal acquisition means. Thus, the target can be polarized with a petard 30 risk zone located at a location other than the signal acquisition zone. Either the target or the polarization and signal acquisition devices may be physically movable to achieve this effect.

6

It is also an object of the invention to provide a computer program product or recording which includes instructions for performing the present method on a magnetic imaging device.

The invention provides considerable advantages. Namely, polarization coding provides information on an object that would otherwise not be available without the addition of gradient fields or encoding by RF pulses in the measurement field. The method can thus both shorten the shooting times and provide additional information about the subject. A particular advantage is that since one or more sample dimensions can be encoded by varying polarization of the sample, the gradient fields over the sample are not required as much as in traditional phase and frequency coding during signal acquisition. Alternatively, or in addition, a time-effective gradient field prior to signal acquisition, as in traditional phase coding, is not required before the signal is collected.

15 In low fields, encoding with RF pulses is difficult because of the low Larmor frequencies and the targeting of the pulse effect is difficult. Polarization coding also solves this problem. The method also reduces the subject's exposure to RF fields and their effects. The effect of RF fields on a subject is measured e.g. SARs (Specific Absorption Rate), the adjusted maximum of which must not be exceeded when shooting live subjects.

20

By polarization coding, it is possible to replace conventional signal averaging, whereby not only the noise effect is reduced, but also additional information is obtained. The polarization coding method can be applied to all polarizations as long as their profile is known. For example, the method does not require linear or homogeneous polarization fields, which makes the method easier to use and makes it very general and well-suited for many purposes.

i oo x Polarization coding can be combined with other coding methods such as Fourier

CL

encoding, so you can take advantage of the best of the various methods. For example, it is possible to encode one dimension with Fourier coding and another with polarization-o coding. The method is applicable not only to MRI but also to magnetic particles (MP, CM ...

in english. magnetic particle) for example by magnetore-laxometry (MRX). Such particles can be used as markers in medicine.

7

The method is suitable for use with one or more reception vessels. For example, using two different polarization fields and one receiving coil provides similar information as using two receiving coils and one polarizing field. For example, we include "coils" as superconducting magnetic field sensors, such as SQUIDs, and high magnetoresistance (GMR) mixed sensors, atomic magnetometers, optical magnetometers, and the like.

An advantage of the invention is that the number of physical receiving coils may be physically limited due to their interconnection or lack of space in the imaging environment. This problem is solved by the invention because the number of physical coils can be reduced without losing information. At the same time, there is also a cost benefit.

The gradient coils required for phase or frequency coding must be implemented very precisely, since the gradients must be linear. An advantage of the invention is that the implementation of various polarization paths, or other equipment providing variable polarization, is not as demanding as long as the shape of the magnetic fields is known (for example, the shape of the fields can be measured after the manufacture of the equipment). Furthermore, the need for power supplies controlling gradient fields can be avoided and, for example, different polarization coils can be driven by a common power supply. In addition, polarization and measurement fields are generally easier to control than rapidly varying gradient fields or RF pulses.

Because polarization coding provides additional information without using gradient fields during measurement, it is easier to combine the method with passive detection of magnetic fields, such as magnetoencephalography (MEG) or magnetocar-o-biography (MKG), which can thus be performed simultaneously with magnetic imaging signaling. Here, for example, the positions o0 of the MRI signals can be coded before the actual signal acquisition, so that simultaneous MEG measurement and acquisition of the magnetization signals are possible. This saves considerable time resources and enables completely new types of measurements, for example in brain research, to be performed over time. The method can also be applied to functional magnetic resonance imaging (fMRI).

CM

MRX has not yet found a valid and well-established way of coding relaxation signals. Typically, particle distribution has been studied either by multiple sensors or by moving one of the 8 sensors. The MPI (Magnetic Particle Imaging) method (B. Gleich, J. Weizenecker, Tomographic Imaging Using Nonlinear Response of Magnetic Particles, Nature 435: 1214-1217, 2005) has not been successfully implemented in practically relevant target volumes. The present polarization coding-5 method allows obtaining additional information about the distribution of the magnetic particles of a target and their properties. For example, with the concept of virtual sensors presented below, new information can be utilized very easily, in the same way that signals from physical sensors are usually processed. By using different polarization fields in support of a limited number and location of physical sensors 10, the object particle distribution and particle properties are more accurately and easily determined.

Throughout this document, the term "magnetic resonance imaging" is used to refer to the measurement of the magnetic properties of particles in a substance and the formation of an image based on measurements. The particles may be atomic nuclei, such as in conventional MRI, or magnetic nanoparticles, such as in MRX, or electrons, as in EPR (Electron paramagnetic resonance imaging).

A "polarizing magnetic field" (or "polarization field") is a magnetic field that shifts an object from a disordered magnetic state (net magnetization substantially zero) or from a state of lesser magnetization (e.g., when magnetization does not completely disappear before a new polarization state is applied). wherein the particles to be measured are at least partially magnetically ordered in the subject (net magnetization is substantially non-zero, i.e. large enough to be measured). The difference in the profiles of the polarizing magnetic fields is essentially that the phase and / or amplitude distribution cm of all the fields used is simply not constant by scaling. Thus, typically at least oo most of the fields used have an inhomogeneous distribution within the target region. It should be noted, however, that among the polarization fields used, there may also be a few homogeneous magnetic fields, for example two homogeneous fields facing each other in a vertical direction, or only one homogeneous field. The forms of the distributions are known computationally o or through experiments or simulations, o

CM

'' Polarization coding '' means a combination of an imaging sequence and imaging technology that utilizes various polarizing magnetic fields to form an image.

9 '' Imaging sequence 'means operations performed sequentially in time which result in polarization of the target and controlled acquisition of the signal in a sufficient number of times so that at least one-dimensional image of the target can be generated from the signals. The sequence may also include steps for performing, for example, slice, phase, frequency, or RF coding, for example, for two- or three-dimensional imaging, and controlling polarization and measurement fields, for example, to perform pre-polarized MRI, as described above.

10 The polarization fields and the measurement field are typically short-term, constant time fields in pre-polarized MRI that have substantially perpendicular components in the sample region. On the other hand, the measuring field may also be continuously on. The coils or sensors used for signal acquisition, in turn, are typically adapted to the resonance frequency (Larmor frequency) of the particles being measured, which typically falls within the RF range (except for very low fields).

Preferred embodiments of the invention and other advantages of the invention will be apparent from the following detailed description, with reference to the accompanying drawings.

Figure IA is a flowchart illustrating the flow of the present method according to one embodiment.

Figure IB is a flowchart illustrating the flow of the present method according to one embodiment (composite coding).

Figure 2 illustrates an exemplary principle of an imaging sequence, Figure 25 illustrates as a graph an error in the present method with different polarization fields, Figure 4 shows phantom voxel values and simulated voxel values using different polarization fields.

Χ Q_

Figure 5 illustrates the effect of various polarization field configurations on the measurement precision r2 30.

LO

oo o o

CM

As discussed above, in the present method, unlike traditional homogeneous first-order polarization, the target is repeatedly polarized mainly by inhomogeneous magnetic fields. Most preferably, the polarization field is initially switched on at the beginning of each round of the imaging sequence to produce magnetization to the target, as is generally the case in pre-polarized MRI. The polarization field is thus switched off during signal acquisition. The measuring field may be connected over the sample either continuously or it may be switched on, for example, after switching off the polarization field before starting the signal generator.

Figure 1A illustrates one embodiment of the invention. After the imaging is started (step 10), the object is magnetized in a location-dependent polarization field (step 11). When the polarization field is switched off, the measuring field is turned on (step 12) 10 (if necessary, the measuring field can also be continuously on). The measuring field is preferably homogeneous, but may also be inhomogeneous. The precession signals produced by the sample in the measurement field are measured using sensors as described above (step 13). During the signal measurement time, different gradient fields can be added to the measurement field as needed. When the signal is collected, the measurement field is turned off (step 14). This is repeated by alternating the profiles of the 15 polarization fields until enough information is obtained about the subject (choice 15).

The amount of information required each time, i.e. the number of iterations K, will always depend on whether coding methods other than polarization coding are used, in how many dimensions the object is represented, and on the desired number of voxels in each dimension. Because different polarization profiles are used with each repeat, the sample is magnetized at different times. Once the variations in the polarization (magnetization) of the sample are known, the mathematics described below can be used to create virtual sensors. Finally, an image is produced from the object using known polarization field information (step 16).

Of course, the same polarization field can also be polarized several times to average the signals and thus improve the signal-to-noise ratio, or if, for example, the same polarization is required for phase and frequency coding, for example.

i oo x Virtual sensors can be utilized in image acquisition, as

CL

is described. On the other hand, if the polarization fields (e.g., in one dimension) are from a function set of r5, for example, wavelets, the image can easily be formed by inversion. A similar technique has traditionally been used, e.g., in RF coding, in which magnetization profiles are formed by modifying RF pulses. Ex. J. B. Weaver, Y. Xu, D. M. Healy, J. R. Driscoll, Wavelet-Encoded MR Imaging, Magnetic Resonance in 11

Medicine 24: 275-287, 1992. According to one embodiment, the polarization fields are shaped according to the Fourier strain.

Figure IB illustrates an embodiment combining polarization coding 5 with conventional Fourier coding. The polarization step 21 occurs as described above, but in this method, one dimension of the object is coded by a phase gradient (step 23) and one dimension by frequency coding (steps 24-26). The sequence is repeated until both the desired number of voxels in the phase coding direction (decision 28) and the polarization coding direction (decision 29) are reached. The image is formed (step 30) by combining the mathematics shown in more detail in the subsequent 10 mins with traditional Fourier coding mathematics.

Of course, after a single polarization, multiple signal echoes can also be used with RF pulses or with multiple phase and frequency coding gradients, as will be understood by one of ordinary skill in the art in conventional imaging techniques.

15

Various polarization fields, i.e., polarization profiles that vary at different times in the target region, can be produced in many different ways. For example, different resistive polarization coils can be used, each with its own power supply or a common power supply, which can be connected at different polarization times to different coils or combinations thereof. It is also possible, for example, to use 20 superconducting polarization fields, which are switched on at different polarization times by various additional fields. Correspondingly, a permanent magnet can produce a polarization field to which various polarization fields are added. Mobile or deformable magnetic field generation means, such as polarization coils or permanent magnets, can also be used. For example, it is also possible, for example, to bring objects, such as metal disks, which influence the magnetic field 25 in a controlled manner at different polarization angles in different ways by varying polarization angles.

i oo x According to one embodiment, the polarization field itself is produced as a standard

CL

as an inhomogeneous field, but the object is moved between different polarization cycles by rotating 30 and / or moving along one or more coordinate axes. Here too, the desired situation is achieved where multiple fields with substantially different location-dependent profiles are formed on the object by fields. The object may thus be placed in a holder which is movable. Also, the measuring coils may move with the movement of the holder, thereby maintaining their relative position with respect to the target.

12

Figure 2 is a plan view (not to scale) of a description system by which the present method may be implemented. The imaging area in which the object to be imaged can be located is denoted by reference numeral 43. A plurality of polarization coils 40 (in this case, 5 to 5 decks) are disposed around the imaging area. Each polarization coil 40 is connected to a power source (not shown) and their currents can be independently controlled. Different polarization profiles can be produced by various combinations of coils. The coil combinations and the currents passing through the coils can be controlled, for example, by computer-assisted or manual-10 operation. Typically, there are at least four polarization coils.

Also in the vicinity of the imaging area are measuring sensors 41. In this example, the measuring sensors 41 are arranged symmetrically and substantially equally around the imaging area 43. Each measuring sensor 41 is connected to the measuring electronics (not shown). Measurement sensor-15 on slots 41 and electronics is treated e.g. signal acquisition and conversion of signals to digital, optionally filtering the signals. Depending on the type of measuring sensor, the electronics can also be used to switch off the sensors, for example, during the effect of a polarization field on the sensor, for example.

Reference numeral 42 denotes coils for producing a measurement field. In this example, the measurement field is generated by a tracked Helmholtz coil pair around the imaging area 43. The current of the measuring coils is preferably controlled by its own power supply (not shown), which can be controlled by computer or manually.

The apparatus typically comprises a central computer or a plurality of computers controlling the hardware for implementing a predetermined imaging sequence, storing the signals collected by the receiving coils, and processing the signals to produce an image of the subject.

X

I do not

CL

The hardware debugging of Figure 2 is also applicable to MRX. However, in this case, the confidence coils 42 of the meter 30 may be omitted or unused.

oo o o c \ i

According to one variation, the imaging area 43 has a stiffened coil (s) (the so-called main polarization coil (s)) which always produce a uniform polarization field over the entire imaging area 43 (similar to the measurement field coils 42 -in different iterations.

The polarization coils and the measuring sensors can be placed very freely in the vicinity of the imaging area and can also be of different sizes. Thus, they need not necessarily be symmetrical as long as a plurality of substantially different polarization field profiles can be produced on the target region and the target region can be detected by at least one sensor. Although this illustrates in more detail only the polarization profiles produced by multiple coils, the same principle applies to the other ways of generating polarization-10 fields listed above.

According to one embodiment, at least some of the polarization fields are local, i.e., such that at least a portion of the object (or its imaging region) is substantially un magnetized. For example, local polarization fields can be generated by small coils. If necessary, for example, another corresponding coil in the second position may be placed in the vicinity of such a polarization coil. This will cause the magnetic field to decrease rapidly, far away. It is also an advantage of local polarization fields that the energy required to produce a magnetic field is lower than, for example, in the case of a polarization field which covers the whole object, but with the same local effect. The advantage obtained can be used, for example, to either amplify the polarization field locally to obtain a larger signal in the measurement step, or to reduce the requirements of, for example, power supplies, so that imaging can be accomplished, for example, with smaller and cheaper equipment.

It is also possible to polarize the sample elsewhere and transfer it to the measurement environment, or to move the polarization and measurement devices and hold the object in place (or to move both). All control and possible transfer of the object or measuring equipment can be done either computer-controlled or without a computer.

X

Signals can be collected by any one or more of known devices for measuring magnetic field or ra 3 ° change, such as coils, SQUIDs, particularly GMR, i.e. high-slip magneto-resistors, mixed-magnetometers, or optical oj magnetometers.

14

The total measurement duration can be shortened by an embodiment in which polarization and measurement (signal acquisition) are performed at least partially simultaneously, but in different parts of the object. This can be accomplished by simultaneously generating a sufficiently strong polarization field at one site and a magnetic field suitable for a measurement field, for example of smaller amplitude and substantially homogeneous in shape. There may be more than one such arrangement. When different field form methods are used alternately, the object can be polarized in other parts of it and, on the other hand, the relaxation of the previous (or preceding) polarization in other parts of the object can be measured.

In this way, it is possible to significantly reduce the imaging times in the pre-polarized MRI, since both polarization and measurement can be performed simultaneously and there is no need to wait for the polarization phase to end before the measurement event is started. Means commonly used in magnetic imaging to encode target signals can be implemented in the measuring range, for example by means of various gradient fields.

On the other hand, simultaneous polarization and signal acquisition is also possible, for example, if the polarization field used in one repetition is sufficiently homogeneous to serve as a measurement field of one of the previous repetitions.

Simultaneous polarization and measurement can also be used to perform, for example, MEG or MKG-20 measurements during MK polarization. For example, when one part of the head is polarized, MEG signals can be simultaneously measured from the other parts of the head. On the other hand, an object can be polarized from other parts of the object and simultaneously measured, for example, MRI and MEG signals from other parts of the object. At present, it is not possible to perform MEG measurement at substantially the same point as polarization because, for example, SQUID sensors, for example, are not capable of signal acquisition in strong magnetic fields. The method is also well suited for example to magnetorelaxometry.

o oo In MRX, on the other hand, the object can be polarized, and where the effect of the polarization field x is different from that of the previous polarization field,

CC

Q_ relaxation after thi field.

<o to ^ 30

LO

g Next, the present polarization coding for multiple signal receiving sensors is approached mathematically:

In magnetic resonance imaging, the signals of multiple sensors as a function of time t can be written to 15 vectors s (t), whose i-th line corresponds to the zrnene sensor signal. Correspondingly, source magnetizations can be written by component into vector m (t). There is a connection between the two vectors s (t) = A m (t) 5 where A is a coupling matrix that describes how the source components are mapped to signals. Assume that K consecutive measurements are performed such that the magnetizations for the t.rme measurement are mk (t) = Ckm {t) where Ck is the transformation matrix. In this case, the signals from the Finnish measurement can be expressed as sk {t) = [ACjfc] m (i)

By combining the signals s * and the matrices AC * by rows it is possible to create a large signal vector 5 'and a generalized coupling matrix A':

f S! (i) \ [ACA

s '(t) -: and A' = i

\ sK (t) J \ ACKJ

15 Now we can write s' (t) = A rm (t)

If the conversion matrices are properly selected,

rank A '> rank A

that is, the matrix A 'has more linearly independent rows than the matrix 20A. Because the degree of the matrix A' is greater than A, s 'and A' can be used to solve the source vector m more accurately than traditionally sm and A: n. Thus, c \ j has been obtained. more information on source magnetization.

00

In order to solve the source magnetization, it is necessary to know the transformation matrix C *. This voice 25 produces a mathematical test of the shapes of the polarization fields, for example, in the following common-co, CO 1a.

1 ^ m 00 o o

Assume that the measurement field is Bm (r ', t) = Bm (r \ t) ez. ex ey and ez are unit vectors in three orthogonal directions. The homogeneous polarization-30 field Bp = Bvex is assumed as the reference point. With this polarization field, the signal equation would become 16 s (f) = A m (t) During the A polarization, the polarization field is B (k) (r r). With this polarization, the signal equation for Finne's measurement is given by 5 »k {t) '·' · '·' · A.Ci.rnU) where Ck is the transformation matrix of the mittath measurement. If the source voxels have a total Nv, the magnetization components corresponding to each voxel are placed in the vector m in succession, and it is assumed that only the x-Jay components of the magnetizations are mapped to signals, C /, is a 2NV x 2Nv block diagonal matrix. Matrix i: s block on / 'f {BP ° (^> ex Spfe) {r'> e? Λ F (fc) = n / / \ {cos ^ (r?) - sm <f> k {rWp ^ -

* l) ^ Sin0fc (r5) COScfyfefy ')) 1 Bjk) (r ^) - ey B {pk \ r'ye: l: I

jq \ Bp Bp / where r'i is the position of the z-voxel, ^ {i \ (rfi ex C0S ^ (r ?;) = || R (fc) r Mil

\\ BP ML

M '] | KV') || W

define the angle between βρ: η and Äp (i) (r'i), || - || xy is the Euclidean norm of the x components of the argumentx and || £ p ° (r ') || £ f t \ Il F 1 II.rv ik {r.}) ----- 15 Bv defines the transform in amplitude. If 11Jfy ^ fy'ijllxy = 0, the angle -φ ^ 'ί) is not specified, but the / s block of the transformation matrix is in accordance with the latest form of the above equation.

i 00 | 20 Using one physical sensor and twice the same polarization field, or using one physical sensor and once two different polarization fields, is (at least approximately)

CD

the same effect on the signal-to-noise ratio of a measurement. However, a particular advantage of the present method is that it provides additional information such as the mathematical

CVJ

addressed. In addition, this is demonstrated by simulations, as can be seen below, for example, in Figure 4 and the related description.

17

It is also conceivable that the measured signals Sk (t) correspond to '' virtual sensors ''. Thus, the present method can reduce the number of sensors, which is an advantage, especially in low field imaging, where 5 expensive superconducting sensors are often used as reception coils. Again, various polarization field profiles can be easily and advantageously produced by resistive coils without significant loss of the measurement signal-to-noise ratio.

The formation of virtual sensors can also be addressed in continuous geometry.

The signal of a 10 nncn sensor as a function of time t in the case of homogeneous polarization can be written in the form sn (t) = j Ln (rf) M (r'J)) dr 'where Ln (f) multiplies the current. the sensor sensitivity at position r 'and M (r', t) contains the source magnetization vector and consists of the components of the magnetization associated with the position r. When the k th polarization-15 th field is such that at source f the source magnetization is of the form Ht (r ') M (r', t) where H /, (r) is a square matrix, the signal of the nth sensor is obtained as 4fc) ( *) = j Ln (rr) == j R} av '> LriF) \ Mir'J) dr' • 'p ,, where Hk is the transposition of the matrix H *. This can be interpreted as the signal of a virtual sensor with a sensitivity at f of H /, (r) '/.' (> '). The signals of these virtual sensors may be processed in much the same way as the signals of the physical sensors.

δ cv i g The performance of the invention has also been demonstrated by simulation results, which are briefly presented below.

i cc

CL

φ Simulation results were generated using the Physikalisch-Technische Bundesanstalt (PTB, Berco

Iin) 304-channel SQUID sensor setup. SQUID sensors are oo § placed inside a flat bottom devar in eight horizontal planes with a diameter of 244

(M

30 mm. The lowest and highest sensor levels are 29 and 169 mm above the bottom of the devar. Refer to stä for more information about the sensor system. Schnabel, M. Burghoff, S. Hartwig, F.

18

Petsche, U. Steinhoff, D. Drung and H. Koch. Sensor configuration for a 304 SQUID vector magnetometer, Neurology and Clinical Neurophysiology, 70, 2004.

The system geometry was implemented on the computer on which the simulations were performed. For Simu-5 plots, a 15x15 voxel digital phantom was placed 40mm below the lowest SQU1D level in the center of the devar. Phantoms were produced using BrainWeb data (http://www.bic.mni.mcgill.ca/brainweb/).

Define the coordinate axes so that the origin is in the middle of the lowest SQUID plane, ez 10 points upwards and ex and ey determine the horizontal plane. Bm = was selected as the measurement field

Bmey, where Bm = 5 μΤ.

The simulations were performed with the following description sequence. Initially, the sample is polarized with the Polarization field Bp (k), after which the field is switched off and the measuring field Bm is turned on. Signals were collected at a sampling rate of 1 kHz with Bm on for 100 ms.

The sensor signals were simulated from phantom voxels, and then white Gaussian noise with standard deviation σ was added to each signal. The magnetization components perpendicular to the measuring field were solved from the above derivative of s' (t) = A2m (t) using 20 generalized coupling matrices, its singular value decomposition, and setting small singular values to zero (regularization) before solving the inverse problem. The magnetization components of the voxels were solved separately for each time point, after which a complex sinusoidal function representing the attenuation was applied to each component. From the parameters of the functions, the amplitudes of the voxel magnetizations were solved.

δ 25

(M

i g The accuracy of the solutions was evaluated by the function i “P 1 - amp £,. | M (ri) |

CL

co where the summations are performed over all voxels, r ', · is Before the voxel position coordinate and M and M represent the magnetizations and the estimated magma 30 net isation used in the simulation, respectively.

The different polarization fields used in the simulation are listed in Table 1.

19

Table 1. Polarization fields used in the simulation (r 'is position coordinate) B B \ -!] = Bpex C Bp1 = Bpe,. Bf = (] + Bpe, D Bp = Bfe, Bp = Bve, E B <'> - Bpe, BP = tl + ^ l) Bper

Bp = Bpe, Bf = {l + Bpe, 5

Figure 3 shows the mean Eamp errors for a 15x15 voxel phantom. Curve labels refer to different polarizations as shown in Table 1. The horizontal axis unit is χ __ σ 10 where Bp is the amplitude of the polarization field, Np is the number of different polarization fields, and σ is the standard deviation of the noise. Defining this, λ increases with the acquisition time (when Np and Bp are attached, σ 1 / Vt where T is the acquisition time).

Figure 4 shows the source phantoms and the resulting solutions, i.e., the amplitudes of voxels for an inactive 15x15 voxel phantom. The signal-to-noise ratio increases from left to right. in the i th column λ = 1010 + 1. Row A shows the source phantoms as such and, in the other rows, the solutions obtained using the different polarization models according to Table 1.

δ

(M

Figure 5 shows the singular values of the generalized coupling matrices A 'for a 0x Wj 15x15 voxel phantom. The values are in descending order so that wo is the maximum singular value of x. The polarizations used are according to Table 1, cc

CL

C/O

The simulation results also show that polarization coding provides more information

LO

o 25 actions to solve source magnetization. It is also noteworthy that using one field polarization field, the information is not sufficient to accurately solve the source magnetizations at any noise level, since some of the singular values of matrix A 'remain zero.

20

It can be seen from Figures 3 and 4 that, within a given measurement time, the source magnetizations can be more precisely solved using polarization coding.

Example 1 (imaging sequence for MRI)

The exemplary description sequence below is intended to assist in understanding and applying the invention in practice. The sequence is designed to combine the polarization coding described herein with Fourier coding. The sequence makes it possible to produce from a subject 10 to produce a three-dimensional image by encoding one dimension by polarization and two other dimensions by phase and frequency coding:

In the measurement process, the following is performed with different polarization fields: 15 fork = ltoKdo for i = 1 to I do a) The kth polarization field is turned on (the sample is magnetized).

b) the kth polarization field is switched off.

(c) The measuring field shall be switched on.

D) Add to the measurement field an x-direction phase coding gradient whose amplitude depends on i.

e) disabling the x-direction phase coding gradient.

f) Adding a y-direction frequency coding gradient to the measurement field during which signal acquisition is performed by at least one receiving coil.

o 25 g) The y-direction frequency coding gradient is turned off.

cm h) Switch off measuring field.

i oo end x end cc

CL

CD

30 Alternatively, the measurement field may be continuously on. When using Fourier coding and 00 o linear gradients, the measurement field should be homogeneous.

C \ l

Finally, a three-dimensional image of the object is generated by the measured signals, the x and y dimensions can be decoded by the signals collected by Fourier transform, the z dimension is obtained 21 decoded by utilizing information generated by various polarization fields and sensitivity profiles of the receiving coils.

Example 2. (mapping sequence to MRX) 5

The present invention is applicable not only to conventional MRI, but also to magneto-magnetic relaxation (MRX), i.e. magnetic particle imaging.

MRX can study the properties of magnetic marker particles, for example, their 10 distributions in the body. In MRX, the sample (= marker particles) is magnetized in a magnetic field. The field is then switched off and the signals due to sample relaxation are measured. The difference with conventional MRI is that no external magnetic field is used during relaxation measurements. The signals collected by the various sensors can be used to determine the locations of particles and their number. By using polarization coding, virtual sensors can be created and the properties of the particles more accurately and easily determined.

An exemplary mapping sequence for MRX is shown below:

The following is performed using different polarization fields.

20 for k = 1 to K do a) The kth polarization field is turned on (sample is magnetized).

b) the kth polarization field is switched off.

c) Measure the relaxation of the sample with one or more sensors, δ 25 end

(M

CNJ

cp co Finally, an image of the particles of the object is formed utilizing the information produced by various polarization x fields and sensors.

0.

C/O

30

LO

o The work that led to this invention has received funding from the European Community's Seventh Framework C \ | program (2007-2013) in accordance with the grant agreement 200859.

Claims (24)

1. Imaging method for magnetic image generation of an object, in which method (a) produces a measurable magnetic polarization (11, 21) in the object (43), (b) polarization of the object (43) is measured (13, 25) by at least one receiver coil, and (c) at least one one-dimensional image of the object (43) is formed (16, 30) on the basis of the measurement, characterized in that steps 10 (a) and (b) are repeated several times by a plurality of polarizing magnetic fields. with substantially different position dependent profiles in the object (43), and the image of the object (43) is formed (16, 30) by means of said plurality of polarizing magnetic fields from a polarized sample on the basis of the measurements taken. Method according to any of the preceding claims, characterized in that in the image design, said plurality of polarizing magnetic fields is used to encode at least one dimension of the object (43) at least partially. Method according to any of the preceding claims, characterized in that disk, phase or frequency coding is used (23, 24) to encode the at least one dimension of the object (43). o i oo 4. A method according to any of the preceding claims, characterized by X 2 using polarizing magnetic fields which are at least substantially non-linearly variable in the at least one dimension in the region of the object (43). in ^. LO 00 o o 25 Method according to any of the preceding claims, characterized in that in step (a) (11,21), each polarization field is switched on, which is switched before the start of the measurement in step (b). Method according to any of the preceding claims, characterized in that the object (43) is polarized and measured at least partially simultaneously, preferably while performing step (a) at a new repeat at the same time as step (b) at a previous repeat, preferably in locally different areas. 5 Method according to any of the preceding claims, characterized in that a measurement (12-14, 22-27) of the magnetization is carried out in a measuring field with an orthogonal component in relation to the respective direction of the polarizing magnetic field or in the used direction. a measurement field which is substantially parallel to the measurement field and the polarization is rotated relative to the measurement field by means of an RF pulse. 10 Method according to any of the preceding claims, characterized in that at least for the time of the step of the magnetization measurement, a substantially homogeneous magnetic field is coupled over the object. Method according to any of the preceding claims, characterized in that the image of the object (43) is formed essentially by determining the magnitude of the source magnetization m from the equation s '(t) = A' m (t) where t is the time. , s 'constitute a number of K signal measurement signals or their mathematical transformations containing signal vector and A' is a coupling matrix which CM contains information on the profile of the polarizing magnetic field applied in the corresponding repetition in the corresponding manner. X CC Method according to any of the preceding claims, characterized in that the polarization and measurement of the object (43) is carried out in a low magnetic field, the polarization LOg ion is preferably less than 100 mT in the field and the measurement is preferably less than 1 mT in the field. the field. Method according to any of the preceding claims, characterized in that in order to achieve the different polarization profiles, a plurality of different polarization coils (40), one or a plurality of moldable or removable polarization coils, one or a plurality of removable permanent magnets and / or one or a plurality of magnetic field processing pieces. Method according to claim 11, characterized in that at least four different polarization coils (40), permanent magnets or magnetic field processing pieces, which are arranged in the vicinity of the object (43) and which can be controlled independently of each other, are used. 10 Method according to any of the preceding claims, characterized in that at least some of the polarizing magnetic fields are relative to the entire image area. A method according to any of the preceding claims, characterized in that said plurality of polarizing magnetic fields cause varying amplitude and / or phase distributions for the polarization in the object 15 (43). The use of a method according to any of the preceding claims in magnetic resonance imaging (MRI), in particular in prepolarized magnetic resonance imaging (MRI). ° 16. The use of a method according to any one of claims 1 to 14 at C \ J and 20 magnetorelaxometry (MRX). oo The use of a method according to any of claims 1-14 in two electron paramagnetization resonance imaging (EPR). CD i ^. LO co § 18. The use of a method according to any of claims 1-14 in (M structural or functional magnetic imaging together with passive magnetic observation, such as magnetocephalography (MEG) or magnetocardiography (MKG), preferably in simultaneous measurement). Apparatus for magnetic image production of an object, said device comprising: means (40) for generating a measurable magnetic polarization in the object (43), - at least one sensor (41) for measuring the polarization generated outside the object ( Means for forming an image of the object (43) on the basis of the measurements, characterized in that the means (10) for generating the measurable magnetic polarization in the object comprise means for providing a plurality of different polarization profiles in the object. Device according to claim 19, characterized in that the means for generating measurable magnetization in the object (43) comprise a plurality of coils (40), each of which is arranged to form a substantially different polarization profile. Apparatus according to claim 19 or 20, characterized in that the means for generating measurable magnetization in the object (43) comprise at least one moldable or removable coil, a permanent magnet or a movable magnetic field processing guide and means for controlling these to form substantially different polarization profiles, o CM c ? Device according to any of claims 19-21, characterized in that it is arranged to be used in pre-polarized magnetic resonance imaging (pre-X polarized MRI). CD CD 1 ^. JG Device according to any of claims 19-21, characterized in that it is arranged to be used in magnetorelaxometry (MRX). 24. Computer software, by means of a computer readable file or readable storage means, containing a program code or by means of computer readable instructions for implementing a method according to any of claims 1-14 by means of a magnetic imaging device. δ CM CM O 00 X IT Q_ CD CD LO CO o o CM