Method and System for Magnetic Resonance Imaging

The present invention relates to a method and system for magnetic resonance imaging (MRI) and, more particularly, to the method and system in which a suppression effect of magnetic resonance (MR) signals from fat within an object being diagnosed is improved by using a pulse sequence including a procedure of applying a prepulse before scan for imaging or exciting selectively protons of water within the object.

In the MR image acquisition, there is a case in which obtaining a so-called MTC (Magnetization Transfer Contrast) is significant from a clinical point of view.

FIG. 1 illustrates the spectrum of protons contained in water, fat and high polymer. As can be seen in FIG. 1, protons contained in water have a resonance frequency of about 64 MHz under a magnetic field of, for example, 1.5 tesla. However, the resonance frequency of protons contained in fat, as well known, shifts toward the low frequency side (the right side in the figure) by 3.5 PPM by chemical shift, and under the above-mentioned magnetic field shifts by about 224 Hz. Meanwhile, protons contained in high polymer have characteristics such that their frequency width is extremely wide.

If a frequency offset by, for example, 500 Hz from the resonance frequency of protons of water is selectively excited before an ordinary imaging sequence is performed, the level of MR signals from the protons of high polymer decreases as indicated by the broken line. Further, the level of the MR signals from the protons contained in water decreases as indicated by the broken line. This is thought to be due to the fact that the protons of water are cross relaxed or exchanged with the protons of high polymer, which fact is already known as an MTC (Magnetization Transfer Contrast) effect.

Use of such an MTC effect makes it possible to obtain an image having a contrast different from that of the prior art in proportion to the ratio at which the high polymer is present. Also, since the level of signals of the parenchyma portion can be advantageously reduced much more than the level of signals of the blood vessel portions, the MTC effect is applied to angiography in which fine blood vessels are extracted.

FIG. 2 illustrates an example of a pulse for selectively exciting protons of high polymer. The pulse shown in this figure is what is commonly called a binomial pulse formed of a plurality of pulse groups such that the pulse length is set at a ratio represented by a binomial distribution and its polarity is alternately reversed. Particularly, the binomial pulse shown therein is such that the pulse length is set at 1:2:1 (hereinafter such a binomial pulse will be abbreviated as a 121 pulse), and the time .tau. between two pulses is short, thus the total time being set at approximately 1 msec by taking the above-mentioned frequency into consideration.

FIG. 3 is a curve representing the magnetization characteristics of the 121 pulse. The vertical axis depicts magnetization Mz, and the horizontal axis depicts the amount of shift from the central frequency f.sub.0. This curve shows the following: protons having a spectrum value near the central frequency are not excited, thus no influences are left in an imaging sequence performed after the binomial pulse is applied. However, protons having a spectrum value which is apart from the central frequency by 500 Hz or more are excited by the binomial pulse, no magnetization is left, and the level of MR signals obtained in a later imaging sequence is reduced. That is, the range of frequencies at which signals can be extracted in a later imaging sequence corresponds to the top portion of the curve of the figure, and the range of frequencies at which signals cannot be extracted corresponds to the bottom portion of the curve of the figure.

Therefore, when the central frequency f.sub.0 of such a binomial pulse is applied where the frequency is made to match the resonance frequency of protons of water, the protons of water are not excited. However, the protons of high polymer are excited in a bottom portion of 500 Hz or above. Thus, it is possible to decrease the level of the MR signals from the high polymer in the imaging sequence performed subsequently to the above. By selectively exciting protons of high polymer beforehand by using a binomial pulse as described above, it is possible to obtain the above-described MTC effect.

Meanwhile, in the MR image acquisition, as fat which is said to have a small clinical significance does not appear in the image, a so-called imaging method capable of suppressing MR signals from fat is often used. There are a number of imaging methods in which MR signals coming from fat is suppressed. In one of them, a method of using a binomial pulse capable of exciting protons in a frequency selecting manner is known.

Since, as described above, the level of signals which appear in a later imaging sequence decreases in the bottom portion of magnetization Mz of the binomial pulse, the binomial pulse is set in such a way that the resonance frequency of protons contained in fat approaches the bottom portion.

FIG. 4 shows a 1331 pulse set in this manner. The time .tau. between the pulses is set at approximately 2.3 msec (longer) so that the resonance frequency of protons contained in water approaches the central frequency f.sub.0 and the resonance frequency of protons contained in fat approaches the bottom portion offset toward the low frequency side by approximately 220 Hz from the central frequency f.sub.0 of the protons contained in fat.

If such a binomial pulse is applied and then a normal imaging sequence is performed, it is possible to form only protons of free water into an image and suppress the level of MR signals from the protons contained in fat.

On one hand, in an MRI system, uniformity of a static magnetic field becomes a key factor to increase quality of an excitation spectrum (i.e., signal to noise ratio and resolution). Thus MRI systems are generally provided with an adjusting means with a shim coil in order to adjust the uniformity of the field. Particularly in medical use, it is preferred to prevent a patient from being placed in the diagnostic space of a magnet for a long time. Therefore, the first-order shimming (gradient shimming) in X-, Y-, and Z-directions, which is simply operated, is carried out to separate resonance frequencies of water and fat in a spectrum.

However, since the length of the entire pulse becomes as much as approximately 7 msec in the 1331 pulse, this makes a repetition time TR longer. In other words, this pulse has the problem that it is difficult to use the signal for an ultra-fast scan and angiography in which the repetition time TR is made short and used.

If, in contrast, the pulse interval .tau. is made short for decreasing the repetition time TR, another problem occurs, for example, the top portion of the graph of magnetization Mz is expanded and fat is also formed into an image, decreasing the suppression effect of MR signals from fat.

Further, the prepulse is liable to be influenced by ununiformity of the static magnetic field, thus decreasing quality of images. Furthermore, since a frequency range distant from the resonance of water is excited (that is, off-resonance), the MTC effect will also be occurred, decreasing a signal to noise ratio correspondingly to it.

On one hand, there is a drawback with respect to the aforementioned first-order shimming. It is impossible for the first-order shimming to achieve complete uniformity of the static magnetic field; there will be still turbulence in higher-order magnetic components of the static magnetic field. This ununiformity of high-order magnetic components frequently cause the curve of water to deviate from the central frequency f.sub.0 of a prepulse (for instance, a binomial pulse of 1331).

FIG. 5 exemplifies a change in frequency as a slicing position changes, which is caused by the ununiformity of the higher-order magnetic components. The degree of deviation of a certain frequency f=f.sub.1 becomes noticeable as the slicing position is distant from the center in the slicing direction. In case that a prepulse is used in multi-slice imaging in order to suppress MR signals from fat, the aforementioned first-order shimming is effective in making the resonance curve of water coincide with the central frequency f.sub.0, as shown in FIG. 6A, for the central slicing plane in the slicing direction, the central slicing plane usually being coincident with an iso-center. However, at slicing positions moved from the center in the slicing direction, the above-said higher-order magnetic components may cause the resonance curve of water to deviate from the central frequency f.sub.0 of a prepulse, as shown in FIG. 6B, with the result that the resonance curves of water and fat move out of the top and bottom portions of the prepulse, respectively. In consequence, when the multi-slice imaging is carried out, there is a problem that image quality alters with changed slicing positions, because slicing planes at the central position and its neighboring positions in the slicing direction have a great suppressing effect of MR signals from fat, while other slicing planes distant from the central slicing position have a less suppressing effect of it.

The present invention has been achieved in view of the above-described circumstance. It is an object of the present invention to provide a method and system for MRI capable of realizing both the MTC effect and the suppression effect of MR signals from fat by using a binomial pulse.

Another object of the present invention is that, when multi-slice imaging is carried out using a pulse sequence including a prepulse for suppressing MR signals from fat, it is possible to avoid a decrease in such suppression effect which results from turbulence in higher-order magnetic components, thus obtaining stable and high quality MR images over a plurality of slicing planes.

Still another object is to, in case of multi-slice imaging, to prevent a repetition time TR from being increased and suppress MR signals from fat, without using a prepulse.

To achieve the above-described object, the method of magnetic resonance imaging in accordance with one aspect of the present invention, comprises the steps of: performing a first sequence including a step for applying a binomial pulse, and performing a second sequence for collecting MR image data, wherein in the binomial pulse application step, the central frequency of the binomial pulse is offset by a predetermined offset amount to the high frequency side with respect to the resonance frequency of protons contained in water and the binomial pulse is applied.

Preferably, the MRI method comprises a step for offsetting the binomial pulse correspondingly to said offset amount. It is preferred that the offset amounts if from 1 ppm to 4 ppm.

The magnetic resonance imaging system in accordance with another aspect of the present invention, comprises gradient magnetic-field application means capable of applying a predetermined gradient magnetic field so as to superimpose it on a predetermined static magnetic field; RF transmitting/receiving means capable of transmitting RF pulses to a subject body and receiving MR signals from the subject body; and control means capable of controlling the gradient magnetic-field application and the RF transmitting/receiving, wherein the control means controls the gradient magnetic-field application means and the RF transmitting/receiving means in accordance with a predetermined pulse sequence, the pulse sequence includes a first sequence for applying a binomial pulse and a second sequence for collecting MR image data, and the first sequence is formed in such a way that the central frequency of the binomial pulse is offset toward the high frequency side by a predetermined offset amount with respect to the resonance frequency of protons contained in water.

In the MRI method and the magnetic resonance imaging system of the present invention, initially, a first sequence including a binomial pulse is performed. Here, the central frequency of the binomial pulse is shifted toward the high frequency side with respect to the resonance frequency of protons contained in water and applied so that protons contained in fat and protons contained in high polymer are excited.

When the binomial pulse is applied in this manner, mainly fat and high polymer are excited in the first sequence while water is not excited. Next, the second sequence such as a field echo method is performed to collect MR Image data.

Since the fat and high polymer are already excited in the first sequence, the level of MR signals obtained in the second sequence is low. The level of MR signals of the high polymer becomes low in the same manner. Meanwhile, since protons contained in water are not excited in the first sequence, strong MR signals can be obtained in the second sequence.

As another aspect of the present invention, there is provided a system for magnetic resonance imaging, in which a prepulse is first applied to an object being examined placed in a static magnetic field for suppressing an MR signal from fat of the object and a pulse sequence for multi-slice imaging of a plurality of slicing planes is then applied to the object, the system comprising: means for performing a first-order shimming, every slicing plane, to separate resonance frequencies of water and fat within the object; means for detecting, every slicing plane, a difference in the resonance frequencies of the water and fat caused by turbulence in higher-older components of the static magnetic field; and means for correcting, every slicing plane, a central frequency of the prepulse in accordance with the detected difference in frequency in each of the slicing planes.

Still another aspect of the present invention, there is provided a system for magnetic resonance imaging, in which proton spins contained within water of an object being examined placed in a static magnetic field are excited for collecting magnetic resonance signals by a pulse sequence, the system comprising: means for applying a 90.degree. RF pulse to the object, said 90.degree. RF pulse selectively exciting only the proton spins of the water; first means for applying a first magnetic gradient field to the object in parallel with application of the 90.degree. RF pulse, the first magnetic gradient pulse having a first strength of magnetic field and magnetically slicing the object at a desired plurality of slicing positions; means for applying a 180.degree. RF pulse to the object at a certain time after application of the 90.degree. RF pulse together with the first magnetic gradient field, the 180.degree. RF pulse refocusing the proton spins excited by the 90.degree. RF pulse; and second means for applying a second magnetic gradient field to the object in parallel with application of the 180.degree. RF pulse, the second magnetic gradient field having a second strength of magnetic field larger than the first magnetic gradient field in strength of magnetic field.

The accompanying drawings which are incorporated in and constitute a part of this specification, illustrate embodiments of the present invention and, together with the description, serves to explain the principles of the present invention. In the drawings:

FIG. 1 is a spectrum view of protons contained in water, fat and high polymer;

FIG. 2 is a graph illustrating a pulse for selectively exciting protons of high polymer;

FIG. 3 is a curve illustrating the magnetization characteristics of a 121 pulse;

FIG. 4 is a curve illustrating a 1331 pulse;

FIG. 5 is a curve showing a deviation in frequency caused by turbulence of higher-order components of a static magnetic field;

FIG. 6A is an illustration showing a matched state between a prepulse and resonance curves of water and fat;

FIG. 6B is an illustration showing a missmatched state between a prepulse and resonance curves of water and fat;

FIG. 7 is a general block diagram of a magnetic resonance imaging system of a first embodiment according to the present invention;

FIG. 8 is a pulse sequence chart illustrating an MR image pickup method of the first embodiment;

FIG. 9 is a graph illustrating a magnetization curve when a binomial pulse if offset by 150 Hz toward the high frequency side;

FIG. 10 indicates experimental data obtained by measuring the intensity of signals output from water and fat;

FIG. 11A is a pulse sequence for shimming according to a second embodiment of the present invention;

FIG. 11B is a pulse sequence for imaging in the second embodiment;

FIG. 12 is a flowchart for performing the pulse sequences in the second embodiment;

FIG. 13 is a spectrum in which the resonance curves of water and fat are separated

FIG. 14 is an illustration explaining calculation of an offset value at every slicing plane;

FIG. 15 is an illustration explaining offset of a prepulse every slicing planes;

FIG. 16 is a pulse sequence showing for imaging according to a third embodiment of the present invention;

FIG. 17 represents an excitation state in which the resonance curve of only water is contained in the excitation range by a 90.degree. RF pulse; and

FIG. 18 represents a relation between the resonance curves of water and fat and the excitation range by a 180.degree. RF pulse.

A first embodiment of the present invention will be explained below with reference to FIGS. 7 to 10.

FIG. 7 shows in a general block diagram the magnetic resonance imaging (MRI) system of this embodiment.

The MRI system shown in FIG. 7 comprises a magnet 1 generally formed into a cylinder shape having a hollow portion therein for placing a patient, and a static power source 2 for supplying required electric current to the magnet 1. When the magnet 1 works, a static magnetic field H.sub.0, being directed in Z-direction along the body axis of the patient, is formed in the central diagnostic space in the hollow portion.

In the hollow portion, there is provided a gradient coil portion 3. The coil portion 3 includes x-coils 3x . . . 3x for generating a reding-out field gradient in X-direction, y-coils 3y . . . 3y for generating a phase-encoding field gradient in Y-direction, and a z-coil 3z for generating a slicing field gradient in Z-direction. These gradient coils are connected, through a gradient power source 4, to a gradient sequencer 5a arranged in a sequencer 5 controlled by a main controller 6.

An RF coil 7 is disposed in the inner space of the magnet 1, whereby the patient is surrounded by the RF coil 7 and the gradient coils 3x . . . 3z. The RF coil 7 is in charge of transmitting and receiving radio frequency (RF) magnetic pulses to and from the patient. The RF coil 7 is also connected, through a transmitted 8T and a receiver 8R, to an RF sequencer 5b of the sequencer 5.

The main controller 6 has a computer for entire control of this MRI system. The sequencer 5 is also provided with a computer which stores programs of pulse sequences such as FE (Field Echo) and fast SE (Spin Echo) methods. In response to a start signal from the main controller 6, the sequencer 5 will supply pulsed currents to the gradient coils 3x . . . 3z and the RF coil 7 in accordance with a specified pulse sequence.

MR signals emitted from the diagnostic portion of the patient will be detected by the RF coil 7 and received by the receiver 8R to send the MR data to a calculation unit 10, where the MR data being formed into image data by reconstruction processing such as Fourier transformation. The image data can be memorized in a memory unit 11 and displayed by a display unit 12. An input device 13 is provided for inputting necessary information from an operator.

Furthermore, in this embodiment, there is disposed a shim coil 14 for the first-order shimming (gradient shimming) in the hollow portion of the magnet 1. Hence the first-order shimming can be done by regulating current supplied to the shim coil 14.

FIG. 8 shows a pulse sequence performed by the sequencer 5 having the gradient sequencer 5a and RF sequencer 5b.

As can be seen in the Figure, the sequencer 5 is designed to perform a first sequence in which the binomial pulse is applied and a second sequence in which MR image data is collected. It is preferable that the second sequence be performed by an FE method as shown in the Figure.

In the first sequence, a 121 pulse P.sub.121 * is applied on condition that the central frequency of the 121 pulse P.sub.121 * is offset by a predetermined offset amount toward the high frequency side. The sign "*" attached to the reference numeral herein shows that the prepulse can be offset in frequency. The amount of offset, which is predetermined by prescan and memorized, for example, is preferably from 1.0 ppm to 4.0 ppm and, more preferably, from 1.5 ppm to 3.0 ppm. In this case, the frequency becomes from 50 Hz to 250 Hz, and from 100 Hz to 200 Hz, respectively under the magnetic field of, for example, 1.5 tesla.

To perform an MR imaging by using the magnetic resonance imaging system of this embodiment, initially, the first sequence including the binomial pulse P.sub.121 * is performed. Thus, offsetting the binomial pulse P.sub.121 * leads to a situation that the protons contained in fat and high polymer are excited.

FIG. 9 is a graph illustrating the magnetization Mz curve when the binomial pulse P.sub.121 * is offset by, for example, 150 Hz, toward the high frequency side.

As can be seen in the Figure, since the vicinity of the resonance frequency of water is a top portion, the protons are not excited. However, since the resonance frequency of protons contained in fat, i.e., near -220 Hz, approaches the bottom portion, the protons are excited by a predetermined amount. The protons contained in high polymer are excited at the bottom portion.

Next, the second sequence is performed to collect MR image data. Since the fat and high polymer are already excited in the first sequence, the level of MR signals obtained in the second sequence is low. The level of signals of the high polymer becomes low in the same manner. Meanwhile, since protons contained in water are not excited in the first sequence, strong MR signals can be obtained in the second sequence.

As described above, in the MR imaging method and the magnetic resonance imaging system of this embodiment, the central frequency of the binomial pulse is offset toward the high frequency side by a predetermined offset amount with respect to the resonance frequency of protons contained in water and the binomial pulse is applied so that the protons contained in fat and high polymer are excited. Therefore, it is possible to greatly reduce the level of MR signals from fat. Therefore, it is possible to form fine blood vessels, which cannot be seen because these are concealed in fat, into an image.

FIG. 10 indicates experimental data obtained by measuring the intensity of signals output from water and fat, and the offset amount is shown as the horizontal axis.

As can be seen in the Figure, when the offset amount was made from 1.0 ppm to 3.8 ppm, the intensity of signals from water becomes from approximately 100% to 50%, and the intensity of signals from fat becomes from approximately 90% to 45%. When the offset amount is made from 1.5 ppm to 3.0 ppm, whereas it is possible to secure the intensity of signals from water to be from approximately 96% to 84%, it is possible to reduce the intensity of signals from fat to be from 84% to 63%.

Also, the MR Imaging method and the magnetic resonance imaging system of this embodiment makes it possible to not only reduce the level of signals from fat, but also reduce the level of signals from high polymer. Thus, the MTC effect can be obtained. In this embodiment, in particular, since the binomial pulse is offset, the frequency band at which high polymer is excited approaches the resonance frequency of water, making it possible to obtain an MTC effect higher than in the prior art in which the binomial pulse is not offset.

Such a suppression effect of MR signals from fat and an MTC effect can be easily obtained by performing the first sequence prior to the ordinary imaging sequence, and there is hardly no need to extend the repetition time (TR).

Therefore, the MR imaging method and the magnetic resonance imaging system of this embodiment are very effective means for a pulse sequence, for example, an ultrafast scan and angiography in which TR is made short and used.

In processes other than angiography, it becomes possible to obtain the MTC effect while suppressing MR signals from fat.

The MR imaging method and the magnetic resonance imaging system of this embodiment have another advantage that it is possible to suppress the MR signals from veins and to intensity signal from arteries.

Although in the above-described embodiment, a 121 pulse is used as being the binomial pulse, the binomial pulse is not limited to this pulse. A 1331 or 14641 pulse may also be used. With such a construction, the degree of separation of water from fat is further improved.

Next, a second embodiment of the present invention will be explained according to FIGS. 11A and 11B to 15. For simplifying or omitting explanation of the embodiments described hereinafter, the same reference numerals as the above first embodiment will be used for the same or equivalent components or processes.

The second embodiment relates to prevention in deterioration of suppression of MR signals from fat in multi-slice imaging.

To accomplish such purpose, an MRI system according to the second embodiment adopts the same hardware as the first embodiment, in which the controller 6, sequencer 5 and calculation unit 10 are designed to carry out cooperatively a series of processes shown in FIG. 12, which is in accordance with FIGS. 11A and 11B.

At the first step 30 in FIG. 12, for each of the X-, Y- and Z-axis directions, the first-order shimming is to be carried out using, for example, a shimming pulse sequences shown in FIG. 11A, thus being able to obtain a uniform static magnetic field Ho. In such shimming processing, obtained are offset values .DELTA.f . . . .DELTA.f in frequency of a prepulse applied to each slicing plane. In this embodiment, a 1331 binomial pulse P.sub.1331 * will be used as the prepulse, as shown in FIG. 11A.

In detail, the first slicing plane, for example, a central plane facing to a designated iso-center in the slicing direction of a three-dimensional region to be sliced is excited by a 90.degree. pulse to obtain an FID (free-induction decay) signal (refer to Steps 30.sub.1 and 30.sub.2). The FID signal is then processed by Fourier transformation, so that obtained is a specrum containing resonance frequencies f.sub.WAT and f.sub.FAT of water and fat as exemplified in FIG. 13 (refer to Step 30.sub.3). Half-value widths W.sub.WAT and W.sub.FAT are each calculated with respect to the resonance curves C.sub.WAT and C.sub.FAT of water and the value etc. of current supplied to the shim coil 14 (refer to FIG. 7) is adjusted in a manner that those half-value widths W.sub.WAT and W.sub.FAT will be minimized (refer to Steps 30.sub.4 to 30.sub.6). Such adjustments enables the resonance curves C.sub.WAT and C.sub.FAT of water and fat to preferably separate to each other.

After completion of such first-order shimming for a certain slicing plane, processing at Step 30.sub.7 will be done, where the difference between the central frequency f.sub.0 of the prepulse P.sub.1331 * and the resonance frequency f.sub.WAT of water is calculated and memorized as an offset value .DELTA.f(=f.sub.0 -f.sub.WAT), assuming that the prepulse P.sub.1331 * of 1331 is set at a frequency that provides that most effective suppression of MR signals from fat. It should be noted that the frequency offset value .DELTA.f=0 at a central slicing plane in the slicing direction, as the central slicing plane (coinciding with the iso-center) is taken as a standard.

The above first-order shimming and calculation of frequency offset values will be carried out for all the slicing planes (refer to Step 30.sub.8). As a result, as exemplified in FIG. 14, the frequency offset value .DELTA.f(=.DELTA.f.sub.1 . . . .DELTA.f.sub.n) is obtained obtained for each of a plurality of slicing planes S.sub.1 . . . S.sub.n.

Next, a multi-slice imaging will be done by applying the prepulse P.sub.1331 * of 1331, as shown in FIG. 11B, before a scan sequence based on an SE (Spin Echo) method (refer to Steps 31 and 32). In the course of this sequence, the frequency offset value of thus calculated at Step 30.sub.7 is added to the prepulse P.sub.1331 * for suppressing MR signals from fat, thus obtaining a corrected prepulse P.sub.1331 * according to the addition (refer to Steps 32.sub.1 and 32.sub.2). Applying the prepulse P.sub.1331 * thus-corrected to an object through the RF coil 7 allows the entire magnetization property curve caused by the prepulse P.sub.1331 * to move toward the resonance curve C.sub.WAT of water on the frequency axis, every slicing plane. In consequence, at each sepctrum of the slicing planes S.sub.1 . . . S.sub.n as shown in FIG. 15 (corresponding to FIG. 14), the central frequency f.sub.0 of the prepulse P.sub.1331 * coincides with the center of the resonance curve C.sub.WAT of water, even when the resonance curve C.sub.WAT and C.sub.FAT in a spectrum are moved every slicing plane because of higher-order turbulence in magnetic components of the static magnetic field H.sub.0.

In this way, it is possible to exclude steadily the conventional drawback that both the specrum and the prepulse deviate from each other as a slicing plane is distant away from the center in the slicing direction. Namely, at any slicing plane, the top and bottom portions in magnetization property of the prepulse P.sub.1331 * are preferably coincident with the resonance curves C.sub.WAT and C.sub.FAT in terms of their frequency positions, thus accomplishing the suppression of MR signals from fat more effectively, obtaining uniform and stable high-quality images over the slicing planes because of a decrease in ununiformity among the suppression effects of MR signals from fat of the slicing planes, and increasing reliability of the MRI system. Moreover, only the shim coil for the first-order shimming is required in this embodiment, thus avoiding the hardware construction of the shim coil from being larger. In addition, a period of preparation time is short, because higher-order shimming is not required.

The prepulse used in the above second embodiment is not limited to the binomial pulse of 1331, but it is possible to adopt other binomial pulses such as a 121 pulse and pulses of a sinc function and a Gaussian function. Also the sequence for collecting image data, which follows the prepulse, is not limited to the aforementioned SE method, but other sequences such as FE and Fast SE methods.

A third embodiment of the present invention will be explained with reference to FIGS. 16 to 18.

An MRI system according to the third embodiment is constructed to avoid use of a prepulse appeared in the previous first and second embodiments and to be able to perform a multi-slice imaging in which a repetition time TR is shorter and MR signals from fat are suppressed.

The MRI system of the present embodiment comprises the sequencer 5 that performs a multi-slice imaging in compliance with a pulse sequence shown in FIG. 16. The pulse sequence shown therein is base on a fast SE method and capable of selecting planes to be sliced within an object and, at the same time, selectively exciting protons of only water in those sliced planes (i.e., most of the protons of fat not excited). Hereinafter, an imaging method using such a pulse sequence shown in FIG. 16 is to be referred to as an "Water-CHASE (Water Chemical Selective Excitation)" method.

In this Water-CHASE method, as shown in FIG. 16, a 90.degree. RF pulse as an excitation pules is applied in parallel with applying a slicing magnetic field gradient G.sub.a =G.sub.a1 in the Z-axis direction. The 90.degree. RF pulse consists of a sinc function whose .pi. numbers are "-4, +1" .tau. at its both sides, thus side lobes at one side (left side in FIG. 16) being designed to be longer and its total pulse length being approximately 20 msec. In this embodiment, in contrast, side lobes at the other side (right side in FIG. 15) are cut to shorten an echo time TE. The number of .pi. may be changed in a range of "-1, +1".pi. to "-10, +10".pi., if necessary.

This longer 90.degree. RF pulse in time than a conventional technique makes its excitation frequency range Rf shorter, as shown in FIG. 17. This excitation frequency range Rf contains only the resonance curve C.sub.WAT of water within each slicing plane and is adjusted to exclude the resonance curve C.sub.FAT of fat caused by the chemical shift phenomenon. During application of the longer 90.degree. RF pulse, the strength of the slicing magnetic gradient G.sub.a1 is set to be a desired value.

After this slicing, a phase-encoding magnetic gradient G.sub.E is applied together with a reading-out magnetic gradient G.sub.R.

At a time since a period of TE/2 have passed from the application of the 90.degree. RF pulse, a 180.degree. RF pulse as a refocusing pulse (having shorter side lobes than the 90.degree. RF pulse) will be applied under another slicing magnetic gradient G.sub.S =G.sub.S2 larger in strength than the previous one;namely, G.sub.S2 >G.sub.S1. Such differences in the two slicing magnetic gradients G.sub.S1 and G.sub.S2 and in frequency ranges of the 90.degree. and 180.degree. RF pulses creates a situation such that a plane sliced at a time when the 180.degree. RF pulse is applied differs in its sliced thickness by one slice from that at a time when the 90.degree. RF pulse is applied.

Then, at a time when TE/2 has elapsed, an echo signal will be collected under application of another reading-out magnetic gradient G.sub.R.

In the same way, another 180.degree. RF pulse is repeatedly applied correspondingly to specified collection numbers of echo signal (at those time, the slicing magnetic gradient G.sub.S =G.sub.S2), and necessary echo data is collected according to the fast SE method.

In this Water-CHASE method, as understood from the above, a longer 90.degree. RF pulse is applied together with a slicing magnetic gradient (G.sub.S (=G.sub.S1), with the result that a desired slicing plane is magnetically selected and spins of protons contained in only water within the selected slicing plane are selectively excited in a manner shown in FIG. 17. The protons of water are always excited by the 90.degree. RF pulse having a narrow frequency band--that is, on-resonance. After this excitation, a 180.degree. Rf pulse is applied together with different strength of a slicing magnetic gradient G.sub.S (=G.sub.S2), so that the excitation position of this time differs by about a slicing thickness compared with that by the 90.degree. RF pulse. Thus, if a plane containing fat is selectively excited by the above 90.degree. RF pulse, another plane will then be excited by the next 180.degree. RF pulse, thereby an echo signal can not be acquired from the fat.

Therefore, the proton spins of water which had been forced to fall on the X-Y plane by the 90.degree. RF pulse are then refocused by the 180.degree. RF pulse, so that an echo signal is collected by the refocusing. In contrast, as the proton spins of fat had not been excited by the 90.degree. RF pulse, even when proton spins of both water and fat which are excited together by a wider excitation frequency range Rf' as shown in FIG. 18, they are only compelled to fall in the minus Z-axis direction. Hence, an echo signal of proton spins in fat will not be collected, thus being able to remarkably suppress echo signals from fat. And the multi-slice imaging is made possible and high-quality images can be obtained.

Furthermore, the selective excitation of proton spins of water at times of slicing eliminates the need for using prepulses, thereby making the repetition time TR shorter than a case where a prepulse is used. It is also possible to provide an echo signal collection that have higher resistance to ununiformity of the static magnetic field than that using binomial pulses. Further, since on-resonance is carried out, no off-resonance, to excite protons of water, the MTC effect is not occurred, thus avoiding a decrease in MR signal values from water and preserving a higher signal to noise ratio corresponding to a T.sub.1 -weighted image.

On one hand, there are J-couplings of spins of protons adjacent to each other among fat molecules, decreasing signals. In a fast SE method, J-modulation gives rise to a problem in which echo signals from fat is increased. In contrast, since almost all the fat protons are not excited in the above Water-CHASE method, there is nothing to do with the J-modulation. As a result, the Water-CHASE method is applicable to the fast SE method.

As a pulse sequence applied to the Water-CHASE method, any SE method including a fast Se method can be used.