Method for operating a magnetic resonance imaging scanner using fast spin-echo technology to simultaneously generate two different contrasts

Abstract

Method for operating a magnetic resonance tomograph for spatially resolved spin resonance measurement of an object, in particular a living object, which is arranged in a static magnetic field B0, whereby an alignment of the object's spins and a longitudinal net magnetization Mz along the direction Z of the static magnetic field results, and wherein a transverse magnetization component Mxy is generated or changed, in particular by generating a spin flip from the Z direction by a desired flip angle, in particular by 90 degrees, by means of at least one high-frequency excitation pulse (1) in spin resonance, wherein spin echo signals from at least one desired volume element of the object are measured using a sequence of high-frequency pulses (4, 5, 7) and switched gradient magnetic fields (2, 3, 6).wherein the at least one volume element is determined by several gradient magnetic fields superimposed at least temporarily on the homogeneous magnetic field B0, wherein the sequence of high-frequency pulses comprises a first (4) and a second high-frequency rephasing pulse (5), in particular a first (4) and second high-frequency rephasing pulse (5) directly after the generation of a transverse magnetization component by means of a high-frequency excitation pulse (1), wherein from the second high-frequency rephasing pulse (5) onwards both a spin echo and a stimulated echo are generated, wherein in addition to a gradient magnetic field (6) in the frequency coding direction, which is switched on during the measurement acquisition to acquire a line of the raw data space, an additional gradient magnetic field (8) in the frequency coding direction is switched on, which generates echo signals or echo signal components newly generated by the high-frequency rephasing pulses (5, 7),in particular, those not attributable to a refocusing of the original spin echo after the first radio frequency rephasing pulse (4) and / or of the original stimulated echo after the second radio frequency rephasing pulse (5), are shifted out of the raw data space or raw data line acquired in a subsequent measurement, characterized in that, during an evolution time (Tau) between the first (4) and second radio frequency rephasing pulse (5), T2*-induced relaxation processes act on a transverse magnetization component present after the first radio frequency rephasing pulse (4), and a magnetization component present in the longitudinal direction after the first radio frequency rephasing pulse (4) remains unaffected, wherein the absolute time between the first (4) and second radio frequency rephasing pulse (5) is extended by the evolution time beyond the duration of an echo interval,which is present between subsequent equidistant high-frequency rephasing pulses (5, 7).

Description

[0001] The invention relates to a method for operating a magnetic resonance tomograph for spatially resolved spin resonance measurement of an object, in particular a living object, which is arranged in a static magnetic field B0, whereby an alignment of the object's spins and a longitudinal net magnetization Mz along the direction Z of the static magnetic field results, and wherein a transverse magnetization component Mxy is generated or changed, in particular by generating a spin flip from the Z direction by a desired flip angle, in particular by 90 degrees, by means of at least one high-frequency excitation pulse in spin resonance, wherein spin echo signals from at least one desired volume element of the object are measured using a sequence of high-frequency pulses and switched gradient magnetic fields.wherein the at least one volume element is determined by several gradient magnetic fields superimposed at least temporarily on the homogeneous magnetic field B0 and the sequence of high-frequency pulses comprises a first and a second high-frequency rephasing pulse, in particular a first and second high-frequency rephasing pulse directly after the generation of a transverse magnetization component by means of a high-frequency excitation pulse, wherein from the second high-frequency rephasing pulse onwards both a spin echo and a stimulated echo are generated, wherein in addition to a gradient magnetic field in the frequency-coding direction, which is switched on during the measurement acquisition to acquire a line of the raw data space, an additional gradient magnetic field in the frequency-coding direction is switched on, which generates echo signals or echo signal components newly generated by the high-frequency rephasing pulses,in particular, those not attributable to a refocusing of the original spin echo after the first radio frequency rephasing pulse and / or of the original stimulated echo after the second radio frequency rephasing pulse (5) are moved out of the raw data space or raw data line acquired in a subsequent measurement.

[0002] Such a method is known, for example, from the publication by NIENDORF, Thoralf: On the application of susceptibility-weighted ultra-fast low-angle RARE experiments in functional MR imaging. Magnetic Resonance in Medicine: An Official Journal of the International Society for Magnetic Resonance in Medicine, 1999, Vol. 41, No. 6, pp. 1189-1198 and from the publications EP 2 645 119 A1 and DE 43 26 902 C1.

[0003] Magnetic resonance imaging (MRI) is a cross-sectional imaging technique that can depict organs or body parts to be examined without the use of X-rays or radioactive substances.

[0004] For example, anatomical images requiring high levels of detail are typically generated using fast spin-echo (FSE) techniques. Following an excitation pulse, a series of multiple high-frequency rephasing pulses is applied. This repeatedly refocuses the initially excited magnetization, and as the signal decays, it is stored in the raw data space (k-space) using repeated spatial encoding (phase encoding). FSE sequences can be used as single-shot or segmented images. Ideally, the high-frequency rephasing pulses are generated as so-called 180-degree pulses and are treated as such in theoretical considerations.

[0005] By selecting appropriate sequence parameters, primarily the echo time (TE), an FSE sequence can generate either T2-weighted (transverse relaxation time of magnetization) or proton density (PD)-weighted images, or be combined with any other known contrast preparation module of MRI. T2-weighted FSE sequences play a major role in the detection of acute tissue infarcts in the brain or heart muscle.

[0006] A modified FSE sequence—for example, in the implementation of displaced UFLARE—imbues the magnetization with sensitivity to microscopic field inhomogeneities (T2* weighting) by means of an additionally inserted evolution time τ. These field inhomogeneities occur at the boundaries of different tissues (susceptibility jumps) or can be physiologically caused—for example, by increased iron concentration in the blood due to hemoglobin or by changes in tissue morphology due to scarring—and lead to an irreversible loss of the phase relationship of the magnetization. Thus, this sequence offers the possibility of examining the oxygen content of blood or the iron content of tissue, or even the detection of myocardial infarctions. T2*-weighted imaging has a wide range of applications: dynamic imaging after contrast administration, functional MRI, or T2* mapping of the myocardium.

[0007] In most applications, a specific, pre-modulated contrast—such as PD, T2, or T2* weighting—is generated for each image. If, as is often the case, several different MRI contrasts are desired, the corresponding images are generated sequentially in separate image acquisitions, also called sequences. This is time-consuming and can also lead to misregistration of image layers.

[0008] Mapping specific tissue parameters, such as T2 or T2*, is of great interest in MRI. A parameter map is calculated from a series of measurements in which the respective parameter (e.g., T2 or T2*) is successively varied. The individual measurements have high spatial resolution, resulting in very precise parameter maps. This allows conclusions to be drawn about the microstructure of the tissue under investigation.

[0009] One disadvantage of this technique is that it is very time-consuming due to the number of measurements and their resolution. In many cases, several parameters are mapped sequentially, further increasing the time factor.

[0010] Therefore, accelerating parameter mapping is of great interest for clinical applications.

[0011] Additionally, the problem of layer misregistration frequently arises when mapping is performed sequentially. As a result, maps with different parameters are not comparable.

[0012] The object of the invention is therefore to provide a method for operating a magnetic resonance imaging (MRI) scanner by means of which, within a sequence of switched radio-frequency pulses and gradient magnetic fields, several types of contrast can be captured in the recorded raw data, in particular T2*-weighted data and another contrast, for example T2-weighted data or weighting according to proton density. According to the invention, superposition of successively generated echoes is to be avoided.

[0013] This task is solved by supplementing a gradient magnetic field in the frequency coding direction, which is switched on during measurement acquisition to capture a line of the raw data space, with an additional gradient magnetic field in the frequency coding direction. This additional field shifts newly generated echo signals or echo components, particularly those not attributable to a refocusing of the magnetization components originally generated after the second high-frequency rephasing pulse, out of the raw data space or raw data line captured in the next measurement acquisition.wherein, during an evolutionary time between the first and second high-frequency rephasing pulse, T2*-induced relaxation processes act on a transverse magnetization component present after the first high-frequency rephasing pulse, and a magnetization component present in the longitudinal direction after the first high-frequency rephasing pulse remains unaffected by this, wherein the absolute time between the first and second high-frequency rephasing pulse is extended by the evolutionary time beyond the duration of an echo interval that exists between subsequent equidistant high-frequency rephasing pulses.

[0014] The subject matter of the invention can, for example, be used in conjunction with the fast spin echo technique (FSE), according to which further equidistant high-frequency rephasing pulses follow the two aforementioned high-frequency rephasing pulses.

[0015] The signal generated by FSE techniques is typically a superposition of spin echoes and stimulated echoes, which are also referred to as SE and STE respectively.

[0016] The theory of FSE technology describes the application of multiple 180-degree radio frequency rephasing pulses following a 90-degree excitation to repeatedly reverse the dephasing of the magnetization component, which was previously tilted into the transverse plane, and thus generate multiple echo signals. In practice, however, pure 180-degree radio frequency rephasing pulses cannot be realized, so that a radio frequency rephasing pulse typically always tilts a portion of the transverse magnetization back from the transverse plane into the longitudinal Z-direction.

[0017] Furthermore, the theory describes the generation of stimulated echoes after excitation of a transverse magnetization by two successive 90-degree radio frequency pulses. The first pulse flips one component of the preceding transverse magnetization back into the Z-direction, leaving another component in the transverse plane. The second pulse flips the Z-component back into the transverse plane with the same direction of action as before, generating a so-called stimulated echo. However, pure 90-degree pulses cannot be generated in practical measurement applications.

[0018] A high-frequency rephasing pulse, often simply referred to as an ALPHA pulse, can therefore be mathematically considered a linear combination of a 180-degree component, a 90-degree component, and a 0-degree component. When applied to a transverse magnetization, it always generates a component that is rotated 180 degrees in the transverse plane and a component of the magnetization that is flipped back in the Z-direction. In FSE technology, SE and STE thus appear superimposed, at least after the second high-frequency rephasing pulse.

[0019] The excitation pulse flips the magnetization initially stored in the longitudinal axis into the transverse plane. The first high-frequency rephasing pulse splits the magnetization. One part remains in the transverse plane and already generates a seq, while the other part is returned to the longitudinal axis and stored there until the next high-frequency rephasing pulse takes effect. This next pulse returns the stored portion of the magnetization to the transverse plane, where it forms a seq.

[0020] Simultaneously, the refocusing of the first SE generates another SE, which is superimposed on the STE. Depending on the number of intervals between the high-frequency rephasing pulses (interecho intervals) during which the echo-forming magnetization was already in the transverse plane, one speaks of odd (E u ) and straight (E gEchoes. Since SEs and STEs overlap, even and odd echoes always mix (interference) and both contribute to the resulting image.

[0021] In the present method, a gradient magnetic field in the frequency coding direction, switched between the first and second high-frequency rephasing pulses, and a gradient magnetic field in the frequency coding direction, switched before the first high-frequency rephasing pulse, initially causes the spin echo and the stimulated echo to be separated in a recorded line of the raw data space from the second interecho interval at the time of measurement acquisition.

[0022] The invention utilizes the fact that the aforementioned gradient magnetic fields in the frequency-coding direction cause different phase shifts to affect the SE and STE before and after the first high-frequency rephasing pulse. These phase shifts are then transferred to subsequent measurement acquisitions because, in the interim, one of the high-frequency rephasing pulses flips a magnetization component back into the Z-direction, thus remaining unaffected by the gradient in the frequency-coding direction after the first high-frequency rephasing pulse, which therefore only affects the transverse component. This results in a separation of the two echo types within a row of the raw data space when subsequent gradients in the frequency-coding direction are applied. This separation is determined by the phase-coding gradient applied during each echo measurement acquisition.

[0023] Preferably, the gradient magnetic field switched in the frequency coding direction before the first high-frequency rephasing pulse causes a phase shift that deviates from 50% of the phase shift, in particular 25% of the phase shift caused by the gradient magnetic field switched in the frequency coding direction between the first and second high-frequency rephasing pulse.

[0024] The 25% effect setting is particularly advantageous here, as it ensures that the data components of both echo types are evenly distributed around the center of a row within the data space, which offers benefits during subsequent evaluation and Fourier transformation. The resulting separation can also be clearly illustrated using the 25% effect of the phase shift of the first frequency encoding gradient (after the excitation pulse): The transverse magnetization prepared with the excitation pulse (initially with zero phase) is first affected by the frequency encoding gradient, which causes a phase shift of 0.25 times that of the subsequent frequency encoding gradient. This magnetization is partially flipped back in the Z-direction by the first high-frequency rephasing pulse. This portion retains its 0.25-fold phase shift. The remaining transverse portion is phase-shifted by 180 degrees by the same pulse, thus changing its phase to -0.25. The frequency encoding gradient following the first high-frequency rephasing pulse therefore captures a line of the raw data space in which the generated spin echo originates at position 0.25, since the prepared phase has converged from -0.25 to zero there, generating the echo.However, the same frequency encoding gradient then further shifts this magnetization component by a phase shift of 0.75, placing it at position 0.75 relative to the abscissa of the raw data space. The next spin echo of this magnetization would therefore occur at position 0.75 of a row in the raw data space, since this phase is then shifted to -0.75 by the 180-degree effect of the subsequent high-frequency rephasing pulse before data acquisition.

[0025] The second high-frequency rephasing pulse returns the magnetization, previously flipped in the Z-direction, to the transverse plane, but with the same direction of action as before, thus also generating an effective 180-degree phase shift. This returned magnetization is then prepared at -0.25 with respect to the abscissa of the raw data space. Using a frequency coding gradient during measurement acquisition, the STE (sharp echo) is then generated at position 0.25 of the selected raw data row, and, as previously mentioned, the SE (sharp echo) at position 0.75. Therefore, the measurement data of both echo types are arranged centrally around the center of the raw data row, i.e., at positions 0.25 and 0.75 of the raw data space with respect to the abscissa of the space.

[0026] According to the invention, it can now be provided that different weightings or contrasts are captured with the two types of echo.

[0027] One implementation can provide that a T2* weighting is captured using the spin echoes.

[0028] According to the invention, it is provided that, during an evolution time between the first and second high-frequency rephasing pulse, T2*-induced dephasing processes are applied to a transverse magnetization component present after the first high-frequency rephasing pulse, whereby a magnetization component present in a longitudinal direction after the first high-frequency rephasing pulse remains unaffected by this, wherein the absolute time between the first and second high-frequency rephasing pulse is extended by the evolution time beyond the interecho interval that exists between subsequent equidistant high-frequency rephasing pulses, in particular 180-degree high-frequency rephasing pulses.

[0029] This allows for measurement series in which different evolution times are set from zero.

[0030] The clear separation of the two echo types generated according to the invention is used to generate different weightings and thus different image contrasts in a suitable preparatory experiment. Inserting an evolution time τ in the first interecho interval leads to a T2* weighting of the SE magnetization.

[0031] Since the portion of the magnetization that will form the STE is stored along the longitudinal axis during this time, it remains unaffected by the T2* weighting and can therefore be assigned a different weighting, e.g., a T2 weighting or proton density weighting, independently of the T2* weighting of the SE magnetization. Thus, the proposed approach allows for the independent contrast weighting of the SE and STE magnetizations.

[0032] According to the invention, only the repeated refocusing of the first co-occurring pair of SE and STE is used for signal generation, i.e., the pair consisting of the second spin echo and the first stimulated echo. Therefore, the spin echo generated after the first high-frequency rephasing pulse can be discarded for measurement purposes.

[0033] A key feature of the invention, independent of the type of weighting chosen for spin echo and stimulated echo, is that stimulated echoes generated by subsequent high-frequency rephasing pulses are pushed out of the acquisition window by additional gradients along the reading direction. This makes it possible to acquire pure SE and pure STE.

[0034] According to the invention, it is provided that, in addition to one or each gradient magnetic field in the frequency coding direction which is switched on during the measurement acquisition for the acquisition of a line of the raw data space, in particular from (including) the gradient magnetic field in the frequency coding direction which is switched on after the second high-frequency rephasing pulse, an additional gradient magnetic field in the frequency coding direction is switched on, which shifts newly generated echo signals or echo components, in particular those not attributable to a refocusing of the magnetization components originally generated after the second and / or first high-frequency rephasing pulse, out of the raw data space or the acquired raw data space line acquired in a subsequent measurement acquisition.

[0035] A gradient magnetic field in the frequency-coding direction, which is switched on during measurement acquisition to capture a line of the raw data space, is generated essentially within the timeframe of an echo interval, i.e., the time between two high-frequency rephasing pulses, particularly 180-degree pulses. Switching on this additional gradient magnetic field preferably means that both gradient magnetic fields are temporally adjacent, and in particular, do not overlap. This adjacentness can preferably occur without a time interval. The switching on of the additional gradient magnetic field also takes place within an echo interval between two high-frequency rephasing pulses.

[0036] It may be provided that successive additional gradient magnetic fields are alternately switched before and after the gradient magnetic field in the frequency coding direction to capture a line of the raw data space, thereby canceling out the effect of two successive additional gradient magnetic fields on the magnetization component remaining transversely after a high-frequency rephasing pulse.

[0037] In connection with the evaluation and a Fourier transformation to be carried out, the invention can further provide that, prior to an evaluation of the measurement values ​​recorded in the raw data space, which include both types of echo, sub-data sets are formed, each comprising only one of the two types of echo.

[0038] Both subsets of data can then be processed separately to generate image representations.

[0039] When processing the raw data, it may be necessary to swap the data components on either side of the row center in some of the raw data spaces, particularly in every second inter-echo interval. Since separating the echo types in successive echoes also swaps the positions of the echo types in successive inter-echo intervals, this can ensure that, for example, one echo type is always to the left of the row center and another type is always to the right. Subsets can then be easily generated by dividing the raw data space in the middle.

[0040] Unless the previously described selection of the first two frequency coding gradients has resulted in an even positioning of the data of both echo types around a respective line center, further training may additionally provide that a formed sub-dataset is recentered in the raw data space before evaluation.

[0041] One example of this implementation is described with reference to the illustrations. Fig. In practical application, an RF excitation pulse1 with a 90-degree effect is shown, which ideally flips the existing original longitudinal magnetization into the transverse XY plane, as is known in the prior art.

[0042] A gradient magnetic field 2 is then switched on to prepare for frequency coding and thus in the reading direction of a raw data line, which is shown below in the Fig. the phase of the transverse magnetization changes. Here, the effect of the gradient magnetic field 2 is chosen to be, for example, 0.25 times that of the subsequent gradient magnetic field 3.

[0043] This is followed in chronological order by a first high-frequency rephasing pulse 4, which flips back a portion of the magnetization in the Z-direction, and whose phase is maintained over time according to the dashed line. A transverse component, however, is shifted by 180 degrees to a negative phase by this pulse, as visualized by the solid line.

[0044] Between the first high-frequency rephasing pulse 4 and the second high-frequency rephasing pulse 5, a gradient magnetic field 3 is applied in the reading direction, i.e., in the frequency encoding direction. During its effect, a first spin echo, which is discarded by measurement, is generated, and the phase of the magnetization generating the spin echo changes positively from, for example, -0.25 to 0.75. The magnetization stored in the Z-direction (dashed line) remains unaffected by this gradient magnetic field 3.

[0045] The second high-frequency rephasing pulse 5 both flips the magnetization from the Z direction back into the transverse plane and phase-shifts it by 180 degrees, and also generates a 180-degree phase shift of the remaining magnetization component within the transverse plane.

[0046] In the following gradient magnetic field 6 for the measurement of a raw data line, both a stimulated echo at position 0.25 and a spin echo at position 0.75 of the raw data line are now generated, as the separated zero crossings of the phase positions illustrate.

[0047] To account for a T2* weighting, an evolution time Tau is inserted here between the first and second high-frequency rephasing pulse, which extends the absolute time interval between the two high-frequency rephasing pulses 4 and 5 beyond the interecho interval, i.e., the time interval between the later high-frequency rephasing pulses 7.

[0048] Since each high-frequency rephasing pulse 5 or 7, in turn, generates magnetization components in the Z-direction due to imperfections, leading to echoes that are not attributable to a refocusing of the magnetizations of the initially generated SE and STE, the invention provides for at least one additional gradient magnetic field 8 in the frequency coding direction before or after each gradient magnetic field 6. Such an additional gradient magnetic field 8 shifts the phase of the components folded back from the Z-direction into the transverse direction so that they do not appear in a recorded line of the raw data space. Successive gradient magnetic fields 8 are connected in such a way that they cancel each other out with respect to their effect on the magnetization remaining in the transverse plane.

[0049] Fig. This further shows that the SE and the STE change their respective positions from one interecho interval to the next. Therefore, before reconstructing the stored raw data, a reordering of the k-space is performed, as described below. Fig. This ensures that only SEs are located in the right half and only STEs in the left half of the acquisition window, or vice versa. Since this mixes even and odd echoes, which can differ in phase, phase correction can be applied preferentially. Both k-space parts can form subsets that can then be Fourier-transformed separately, resulting in two images. The image reconstructed from SEs is T2* weighted, while the other remains T2 or proton density weighted.

[0050] Since the proposed approach generates two images from one FSE sequence, this solution can be called a 2-in-1 FSE technique.

[0051] The described preparation module 90°-α-τ-α, which only weights a selected part of the magnetization T2*, is not limited to combination with FSE techniques. It can also be used as a preparation experiment in conjunction with other imaging techniques such as echo planar imaging (EPI), gradient and spin echo (GRASE), or gradient echo techniques.

[0052] It is also conceptually feasible to use non-Cartesian phase coding, such as the implementation of a radial k-space trajectory. Additionally, the initial excitation pulse can be replaced by other spin-preparation modules that generate transverse magnetization.

[0053] Conceptually, it is also possible to introduce a different preparation for the SE component instead of the T2* weighting using the evolution time τ. The use of diffusion gradients in the first interecho interval is of particular interest here. The proposed approach is not limited to imaging but is also applicable to NMR and MR-based spectroscopy, including spectroscopic imaging techniques.

[0054] As a further embodiment, the invention can be used to simultaneously generate a proton density weighted (PD) and a T2* weighted image of an object, e.g. the brain.

[0055] For example, an evolution time τ = 15 ms is inserted in the first interecho interval. Fig. shows two resulting consecutive raw data rows and illustrates the different weighting of the SEs and STEs as well as the changing of their respective positions.

[0056] After rearranging all k-space lines (see Fig. ), both halves of k-space are reconstructed separately. A Fourier transform of each half yields a separate image.

[0057] Acquiring a high-resolution brain slice using FSE technology takes approximately 5 minutes; therefore, acquiring a PD and a T2* weighted image using standard techniques takes a total of 10 minutes. With this invention, the data acquisition time is reduced to 5 minutes.

[0058] The proposed technique is not limited to brain imaging but can be extended to other target regions. Imaging of the heart is of particular interest in this regard.

Claims

[1] Method for operating a magnetic resonance tomograph for spatially resolved spin resonance measurement of an object, in particular a living object, which is arranged in a static magnetic field B0, whereby an alignment of the spins of the object and a longitudinal net magnetization Mz along the direction Z of the static magnetic field results and wherein a transverse magnetization component Mxy is generated or changed, in particular by generating a spin flip from the Z direction by a desired flip angle, in particular by 90 degrees by means of at least one high-frequency excitation pulse (1) in spin resonance, wherein spin echo signals from at least one desired volume element of the object are measured using a sequence of high-frequency pulses (4, 5, 7) and switched gradient magnetic fields (2, 3, 6),wherein the at least one volume element is determined by several gradient magnetic fields superimposed at least temporarily on the homogeneous magnetic field B0, wherein the sequence of high-frequency pulses comprises a first (4) and a second high-frequency rephasing pulse (5), in particular a first (4) and second high-frequency rephasing pulse (5) directly after the generation of a transverse magnetization component by means of a high-frequency excitation pulse (1), wherein from the second high-frequency rephasing pulse (5) onwards both a spin echo and a stimulated echo are generated, wherein in addition to a gradient magnetic field (6) in the frequency coding direction, which is switched on during the measurement acquisition to acquire a line of the raw data space, an additional gradient magnetic field (8) in the frequency coding direction is switched on, which generates echo signals or echo signal components newly generated by the high-frequency rephasing pulses (5, 7),in particular, those not attributable to a refocusing of the original spin echo after the first radio frequency rephasing pulse (4) and / or of the original stimulated echo after the second radio frequency rephasing pulse (5) are moved out of the raw data space or raw data line acquired in a subsequent measurement. characterized by, that in an evolution time (Tau) between the first (4) and second high-frequency rephasing pulse (5) T2*-induced relaxation processes act on a transverse magnetization component present after the first high-frequency rephasing pulse (4) and a magnetization component present in the longitudinal direction after the first high-frequency rephasing pulse (4) remains unaffected by this, wherein the absolute time between the first (4) and second high-frequency rephasing pulse (5) is extended by the evolution time beyond the duration of an echo interval that exists between subsequent equidistant high-frequency rephasing pulses (5, 7). [2] Method according to claim 1, characterized by, that successive additional gradient magnetic fields (8) are alternately switched before and after the gradient magnetic field (6) in the frequency coding direction to capture a line of the raw data space, in particular wherein two successive additional gradient magnetic fields (8) cancel each other out in their effect on the magnetization component remaining transversely after a high-frequency rephasing pulse (5, 7). [3] Method according to any one of the preceding claims, characterized by , that the high-frequency rephasing pulses (7) following the first (4) and second high-frequency rephasing pulse (5) are set as 180-degree pulses. [4] Method according to any one of the preceding claims, characterized by , that the first (4) and second high-frequency rephasing pulse (5) are selected identically, in particular each causing a magnetization rotation of less than 180 degrees and preferably not equal to 90 degrees. [5] Method according to any one of the preceding claims, characterized by , that before evaluating the measurement values ​​stored in the raw data space, which include both echo types, sub-datasets are created, each containing only one of the two echo types. [6] Method according to claim 5, characterized by , that to form the sub-datasets the raw data space is halved at the ordinate and either before the halving a swap of the data parts on both sides of the row center takes place in some of the rows of the raw data space or after the halving some of the rows of the sub-datasets are swapped between the sub-datasets. [7] Method according to claim 5 or 6, characterized by , that each generated sub-dataset is recentered in the raw data space before evaluation. [8] Method according to any one of the preceding claims, characterized by, that each pair of spin echo and stimulated echo at successive echo intervals is encoded into a different row of the raw data space by different choice of phase encoding gradient, in particular where the raw data space is filled starting from the ordinate origin alternately to both sides of the ordinate origin and to the edge of the raw data space in a continuously increasing manner.

Images