WO2009024898A2

WO2009024898A2 - Magnetic resonance imaging with dynamically optimized temporal resolution

Info

Publication number: WO2009024898A2
Application number: PCT/IB2008/053234
Authority: WO
Inventors: Daniel R. Elgort
Original assignee: Koninklijke Philips Electronics, N.V.
Priority date: 2007-08-22
Filing date: 2008-08-12
Publication date: 2009-02-26
Also published as: WO2009024898A3

Abstract

A magnetic resonance scanner (10, 16) is configured to acquire a succession of magnetic resonance images of a subject at a frame rate and to measure one-dimensional projections passing through the center of k-space during the acquisition. A frame rate analyzer (30) is configured to determine a maximum temporal frequency from the measured one-dimensional projections. A controller (12) is configured to adjust the frame rate based on the determined maximum temporal frequency.

Description

MAGNETIC RESONANCE IMAGING WITH DYNAMICALLY OPTIMIZED

TEMPORAL RESOLUTION

DESCRIPTION

The following relates to the magnetic resonance arts. The following finds illustrative application to magnetic resonance imaging, and is described with particular reference thereto. However, the following will find further application in magnetic resonance applications generally.

In dynamic or "real time" diagnostic and therapeutic clinical applications, magnetic resonance imaging is used to image moving anatomical structures, structures affected by cardiac cycling or respiration, inflow of an administered magnetic contrast agent bolus, interventional procedures entailing insertion of a catheter or other instrument, or other subjects that are undergoing change over time. Under these circumstances it is desirable to acquire images at a high frame rate so that the motion or other change over time is depicted accurately.

For "real-time" imaging, the frame rate is typically set to the highest value practically achievable by the magnetic resonance scanner. This ensures that relevant information is not lost due to dynamic artifacts. Additionally, there is concern that if the frame rate is too slow, higher temporal frequency components may be aliased to lower frequencies, such that not only is the high frequency information lost but the lower frequency content is distorted by the aliased content.

However, increasing the frame rate typically involves tradeoffs in other imaging parameters, and may result in reduced spatial resolution, lower signal-to-noise ratio (SNR), or otherwise degraded image quality. Other tradeoffs may entail reducing safety margins, for example by applying radio frequency energy pulses at higher energy and increasing the SAR exposure of the subject.

In accordance with one aspect, an imaging method is disclosed, comprising: acquiring a succession of magnetic resonance images of a subject at a selected frame rate; during the acquiring, measuring a plurality of one dimensional projections passing through the center of k- space; determining a maximum temporal frequency from the measured one-dimensional projections; and analyzing the selected frame rate based on the maximum temporal frequency.

In accordance with other aspects, a storage medium is disclosed that is encoded with instructions executable to control a magnetic resonance scanner to perform the method of the preceding paragraph, and a magnetic resonance system is disclosed including means for performing the method of the preceding paragraph.

In accordance with another aspect, an imaging apparatus is disclosed, comprising: a magnetic resonance scanner configured to acquire a succession of magnetic resonance images of a subject at a frame rate and to measure one dimensional projections passing through the center of k- space during the acquisition; a frame rate analyzer configured to determine a maximum temporal frequency from the measured one-dimensional projections; and a controller configured to adjust the frame rate based on the determined maximum temporal frequency.

In accordance with another aspect, an imaging method is disclosed, comprising: acquiring a succession of magnetic resonance images of a subject at a selected frame rate; during the acquiring, measuring a plurality of one dimensional projections passing through the center of k-space; and determining a minimum frame rate equal to about twice a maximum temporal frequency of the plurality of one-dimensional projections. One advantage resides in enhanced image quality in magnetic resonance imaging of dynamic subjects without concomitant loss of temporal resolution.

Another advantage resides in higher resolution images acquired by magnetic resonance imaging without concomitant loss of temporal resolution.

Another advantage resides in enabling dynamic optimization of frame rate in magnetic resonance imaging of dynamic subjects.

Still further advantages of the present invention will be appreciated to those of ordinary skill in the art upon reading and understand the following detailed description.

The drawings are only for purposes of illustrating the preferred embodiments, and are not to be construed as limiting the invention. FIGURE 1 diagrammatically shows a magnetic resonance system for performing imaging of a dynamic subject at a dynamically optimized frame rate. FIGURE 2 diagrammatically shows a timing sequence for acquiring a succession of magnetic resonance images of a subject at a frame rate with one-dimensional projections passing through the center of k-space measured at time gaps between image acquisitions. FIGURE 3 diagrammatically shows the timing sequence of FIGURE 2 modified in that some one-dimensional projections measured immediately adjacent a magnetic resonance image acquisition overlap with and contribute to the immediately adjacent image acquisition.

FIGURE 4 diagrammatically shows a flow chart for acquisition and processing of one set of one-dimensional projections acquired during a time gap in an acquisition of a succession of magnetic resonance images.

It is recognized herein that a slower frame rate than the highest possible frame rate is typically sufficient to provide acceptable temporal resolution. It is further recognized that it is typically advantageous to use the slowest frame rate sufficient to provide acceptable temporal resolution, since using a higher frame rate generally degrades image quality. It is further recognized that it is advantageous to dynamically adjust the frame rate to be optimal for the temporal variation in the imaging region, which temporal variation may itself vary with time. For example, during insertion of a catheter the frame rate is optimally slow when the catheter is moved slowly during precision alignment operations. The slow frame rate provides high image quality which may be useful to the physician so as to promote accurate positioning. The frame rate is optimally increased during operations in which the catheter is moved more rapidly. The apparatuses and methods disclosed herein enable dynamic or "real time" imaging in which the frame rate is dynamically adjusted based on the rate of motion or other dynamic activity (e.g., blood flow, magnetic contrast agent influx, or so forth).

With reference to FIGURE 1, an imaging system includes a magnetic resonance scanner 10. In some suitable embodiments, the scanner 10 includes components (not shown) such as a main magnet for defining a main (Bo) magnetic field, a plurality of magnetic field gradient coils for superimposing selected magnetic field gradients on the main (Bo) magnetic field, and one or more radio frequency coils for exciting and detecting magnetic resonance. The main magnet and magnetic field gradient coils can be arranged as concentric cylindrical components, or can be arranged as two sub-units disposed above and below (or otherwise on opposite sides of) a subject, respectively, or can be otherwise configured. Some suitable magnetic resonance scanners include the Intera , Achieva , and Panorama^™ magnetic resonance scanners available from Philips Medical Systems, Eindhoven, the Netherlands.

In an illustrative imaging application, a magnetic resonance controller 12 operates the scanner 10 to cause the magnetic field gradient coils to operate to spatially limit or encode magnetic resonance excitations and/or the received magnetic resonance signals, which are stored in a data buffer 14. In non-imaging applications, the magnetic resonance scanner may acquire, for example, magnetic resonance spectra or one-dimensional magnetic resonance projections that are stored in the data buffer 14. When the acquired data is spatially encoded imaging data, a reconstruction processor 16 applies a reconstruction algorithm commensurate with the spatial encoding so as to generate a reconstructed image that is stored in an images memory 18. The image may in general be a two-dimensional image (e.g., a single slice) or a three-dimensional image (e.g., a volume image).

If the imaged subject is dynamic, that is, has a characteristic imagable by the scanner 10 that varies with time, then the controller 12 suitably operates the scanner 10 including the reconstruction processor 16 to acquire a succession of magnetic resonance images of the subject at a selected frame rate, and these images are suitably displayed on a graphical user interface 20 as a CINE sequence, or as individual images that a radiologist or other medical person can selectively review. The dynamic characteristic can be, for example: (i) a moving organ such as the heart or lungs; (ii) an interventional instrument such as a catheter or biopsy needle being used to perform an interventional procedure; (iii) a magnetic contrast agent injected into the subject and inflowing into an organ through the bloodstream or another fluid pathway; (iv) blood flow; and so forth. In some embodiments, the acquired succession of magnetic resonance images are processed by the user interface 20 or by another processor or controller to extract useful information such as a cycling rate of a cyclically moving organ (e.g., the cardiac cycle rate, or a respiratory rate), a quantitative contrast agent inflow time measure, or so forth. As another illustrative application, the succession of magnetic resonance images can be elastically spatially registered to remove the effect of undesired dynamic effects, such as undesired movement of a medical patient being imaged.

It will be appreciated that, in order to accurately image or characterize a temporally varying feature of the subject, the frame rate at which the succession of magnetic resonance images is acquired should be fast enough to capture the temporal variation. For example, based on the Nyquist criterion, if the frame rate is at least twice the maximum temporal frequency, it follows that the succession of magnetic resonance images can capture all temporal frequencies including the maximum temporal frequency. However, if the frame rate is less than twice the maximum temporal frequency, then the higher frequencies may be lost or aliased to the lower temporal frequencies. On the other hand, employing a higher frame rate than that dictated by the Nyquist criterion or another selected criterion is also disadvantageous since a tradeoff typically exists between frame rate and other imaging parameters of value such as image quality-related parameters (for example, resolution or k-space sampling density or signal-to-noise ratio (SNR)) or safety-related parameters (for example, SAR or applied radio frequency energy).

Accordingly, a frame rate analyzer 30 is configured to determine a maximum temporal frequency from measured one-dimensional projections that pass through the center of k-space. Advantageously, a one-dimensional projection that passes through the center of k-space represents a projection through the entire image space. The temporal variation of acquired one-dimensional projections therefore contains the entire temporal variation of the image space. However, a one-dimensional projection representation of the image space can be acquired much more quickly than the image itself.

It is to be understood that the specification of a projection "passing through the center of k-space" or the like are to be broadly construed as encompassing projections that pass either through the center of k-space or proximate to, but not precisely through, the center of k-space, such projections also having the characteristic of being at least an approximate representation of a projection through the entire image space.

In an illustrative embodiment, the magnetic resonance controller 12 controls the magnetic resonance scanner 10 to measure a plurality of one-dimensional projections each passing through the center of k-space. The illustrated frame rate analyzer 30 includes a Fourier transform processor, such as an illustrated fast Fourier transform (FFT) processor 32 which applies a one-dimensional FFT to each measured one-dimensional projection. The Fourier transform converts the each one-dimensional projection into a one-dimensional array of spatial frequency components. A correlator 34 processes the FFT -processed one-dimensional projections to generate a temporal frequency power spectrum 36. The maximum temporal frequency is then the highest temporal frequency for which the temporal frequency power spectrum 36 has a value above a selected threshold. A minimum frame rate selector 40 then selects the minimum frame rate to be about twice the maximum temporal frequency. For example, in consideration of the Nyquist criterion, the minimum frame rate selector 40 in some embodiments suitably selects the minimum frame rate to be about twice the maximum temporal frequency. In some embodiments, the FFT 32 or other Fourier transform processor is omitted, and the correlator 34 is applied directly to the one-dimensional projection measurements to produce a temporal frequency representation from which the maximum temporal frequency is obtained. An advantage of including the FFT 32 or another Fourier transform is that the processing can optionally be limited to a range of spatial frequencies. For example, low spatial frequencies tend to correspond to large-scale spatial features while high spatial frequencies correspond to spatial detailing or small-scale spatial features. Thus, in some embodiments the output of the FFT processor 32 is truncated at a spatial scale truncation threshold and only spatial frequency components below the truncation threshold are retained. The truncated spatial frequency components are then input to the correlator 34, so that the resulting temporal frequency power spectrum 36 includes only temporal frequency components pertaining to features larger than the spatial scale selected by the truncation threshold. Thus, for example, if the motion of interest is the cardiac cycling, then by truncating at a spatial scale truncation threshold that excludes spatial frequency components substantially finer than the size of the heart, the resulting temporal frequency power spectrum 36 excludes temporal frequency contributions caused by motion of features that are substantially smaller than the heart.

The minimum frame rate output by the frame rate analyzer 30 can be used in various ways. For example, the minimum frame rate may be input to the magnetic resonance controller 12, which adjusts the frame rate of the acquisition of the succession of magnetic resonance images to comport with the minimum frame rate. Additionally or alternatively, a frame rate alarm 42 is optionally configured to provide a perceptible warning, such as a visual flashing indicator on a display of the user interface 20 or an audible alarm or a combination thereof, if the minimum frame rate is greater than the frame rate being used in the acquisition of the succession of magnetic resonance images. This alarm warns the radiologist or other operator that the frame rate is below the minimum frame rate indicated by the Nyquist criterion (or another minimum frame rate criterion) for capturing all temporal frequencies including the maximum temporal frequency.

With reference to FIGURE 2, the one-dimensional projections are suitably measured in time gaps between image acquisitions. In some embodiments, a set of one-dimensional projections are acquired between each image acquisition. In other embodiments, two, three, or more images are acquired in succession without interruption, followed by acquisition of a set of one-dimensional projections. In other embodiments, projection data measurements are interleaved within an image acquisition, for example by measuring extra data lines through the center of k-space during acquisition of a volume image.

Each set of one-dimensional projections includes a plurality of one-dimensional projections each passing through the center of k-space. The number of one-dimensional projection measurements should be large enough to enable determination of the power spectrum. The number of one-dimensional projection measurements is in some embodiments less than fifteen projection measurements, and is more preferably less ten projection measurements, and is still more preferably between 5 and 10 projection measurements. The rate at which the projection measurements are acquired should satisfy the Nyquist criterion for the highest temporal frequency expected to be observed. For example, if the dynamic characteristic is cardiac cycling, then the highest temporal frequency expected to be observed is on the order of the highest expected pulse rate. Since measurement of a one-dimensional projection is rapid, e.g. of order a few milliseconds, and is substantially more rapid than the acquisition of an image, the Nyquist criterion for the highest temporal frequency expected to be observed is readily satisfied for typical application such as monitoring the maximum temporal frequency of organ cycling (e.g., cardiac cycling or respiration), interventional instrument manipulation, magnetic contrast agent influx, blood flow, or so forth. A maximum temporal frequency and corresponding minimum frame rate is determined from each set of one-dimensional projection measurements. By interspersing measurement sets of one-dimensional projections at a plurality of time gaps in the acquisition of successive magnetic resonance images, the frame rate analyzer 30 can monitor the maximum temporal frequency and minimum frame rate in approximately real time, and the controller 12 can adjust the frame rate in approximately real-time. For example, if a set of one-dimensional projection measurements is acquired after every ten image acquisitions, then the controller 12 can make a frame rate adjustment after every ten image acquisitions. Thus, for example, if the application is monitoring a manipulation of an inteventional instrument during an interventional procedure, the controller 12 can speed up the frame rate during intervals in which the instrument manipulation is relatively rapid, such as when the physician is in the process of inserting the interventional instrument, and can slow down the frame rate during intervals in which the instrument manipulation is relatively slow, such as when the physician is carefully positioning an instrument tip respective to a critical anatomical feature. Advantageously, the magnetic resonance controller 12 can also adjust other imaging parameters such as image-quality related parameters or safety-related parameters in approximately real time as well. Thus, in the immediately preceding example as the physician reaches the critical time at which the instrument tip is positioned respective to the critical anatomical feature, the frame rate slows down due to the relative lack of motion reflected in the lower maximum temporal frequency, and the controller 12 simultaneously can increase one or more image quality-related parameters such as image resolution. Thus, the resolution is reduced but the frame rate increased as during the instrument insertion stage which is desirable since the rapid insertion motion is effectively tracked and high resolution is unnecessary during the insertion. During the critical tip positioning stage, the resolution is advantageously high but the frame rate is slow, which is acceptable since the physician is slowly and carefully positioning the tip and so there is no high-speed motion to be tracked. With reference to FIGURE 3, in some embodiments one or more of the projection measurements can be integrated with or overlap the immediately preceding or succeeding image acquisition. In the example of FIGURE 3, the first projection measurement of each set of projection measurements also provides data for the immediately preceding image acquisition, while the last projection measurement of each set of projection measurements also provides data for the immediately succeeding image acquisition. This overlapping advantageously further increases overall acquisition speed. It should be noted that the acquisition of the successive magnetic resonance images optionally employs a steady state imaging technique in which steady-state maintenance radio frequency pulses are applied to maintain a steady state magnetic resonance. In these embodiments, the timing of the RF excitation pulses can be configured such that the steady-state within the imaging volume is not disturbed.

With reference to FIGURE 4, a flow chart of an illustrative method 50 for measuring and processing the one-dimensional projections to obtain a maximum temporal frequency is described. In a measurement operation 52, a plurality of one-dimensional projections, each passing through the center of k-space, are measured. As noted previously, the measurement operation 52 can be substantially disposed in a time gap between image acquisitions, and optionally overlaps one or both of the immediately preceding and immediately succeeding image acquisitions. The number of acquired projections is optionally less than 15, and more preferably less than 10, and still more preferably between 5 and 10 projections. Fifteen or more projections are optionally measured, with the number of measured projections selected to provide sufficient samples to computer the temporal frequency power spectrum without unduly lengthening the acquisition of the succession of magnetic resonance images. In some cases, it may be advantageous to collect two or three sets of orthogonal projections so as to enhance sensitivity to motion in multiple directions. It may also be advantageous to orient the projections such that they are parallel to the direction which the motion in the image is anticipated.

In an optional transform operation 54, an FFT or other Fourier- type transformation is applied to each measured projection to convert it to a one-dimensional array of spatial frequency data. As noted previously, the transform operation 54 is optionally omitted. If the transform operation 54 is included then in a further optional operation 56 a spatial filter can be applied to remove high frequency spatial components so that the resultant temporal frequency power spectrum does not include extraneous components from high frequency motion or other temporal activity of spatially small features that are not of interest.

In a correlation operation 58, the one-dimensional projections, optionally after transformation and optionally after spatial filtering, are processed by a correlator to generate a temporal frequency power spectrum. This spectrum is bandlimited by the Nyquist criterion to temporal frequencies at or below one-half of the projection measurement (i.e., sampling) frequency. Thus, the measurement operation 52 should measure projections at a rate more than twice as fast as the highest temporal frequency that is expected to occur. As noted previously, this is not a stringent design parameter since the projections can be measured much more rapidly than images can be acquired. In an identification operation 60, the maximum temporal frequency is identified. In suitable embodiments, the identification operation 60 identifies the maximum temporal frequency as the highest temporal frequency in the temporal frequency power spectrum that is above a selected threshold. The threshold is selected to remove temporal frequencies at the noise level, and in some embodiments the threshold is a percentage of the highest value in the temporal frequency power spectrum.

In a minimum frame rate identification operation a minimum frame rate is selected to be about twice the maximum temporal frequency. In some embodiments the minimum frame rate is selected to be twice the maximum temporal frequency in precise accordance with the Nyquist criterion. In other embodiments, a safety margin may be built in, for example by selecting the minimum frame rate to be 2.1 times the maximum temporal frequency. On the other hand, in some embodiments it is contemplated to identify the minimum frame rate as a rate somewhat below twice the maximum temporal frequency in order to trade off some loss or distortion of high temporal frequency data for higher spatial resolution, a higher safety factor, or another competing objective. Although not shown in FIGURE 4, the minimum frame rate can be used in various ways, such as by the controller 12 of FIGURE 1 to set the frame rate for image acquisitions performed after the minimum frame rate identification operation 62, or by the frame rate alarm 42 of FIGURE 1 in order to provide the perceptible warning if the frame rate being used for image acquisition is below the minimum frame rate, or can be tagged to the stored images along with the acquisition frame rate as metadata to provide a record of whether the images were acquired at a sufficiently fast frame rate, or so forth.

As noted previously, the measurement operation 52 can be performed in a time gap between image acquisitions, and optionally has some overlap with one or both immediately adjacent image acquisitions. However, the computational operations 54, 56, 58, 60, 62 can be performed concurrently with the image acquisition. Moreover, for dynamic imaging it is to be appreciated that the method 50 can be repeated occasionally with the measurement operation 52 repeated at successive time gaps during the acquisition of a succession of magnetic resonance images of the subject (for example, the measurement operation 52 can be performed during a short time gap inserted after every 10-20 images have been acquired), and the frame rate used for the acquisition adjusted in approximately real time to comply with the minimum frame rate. The processing components disclosed herein, including for example the controller 12, the reconstruction processor 16, and the frame rate analyzer 30, can be embodied in various ways. For example, these components can be implemented as one or more suitably programmed digital processors or controllers, or as application-specific integrated circuitry (ASIC), or as a digital computer running suitable software, or as any of various combinations thereof. The disclosed components can be combined or separated in various ways. For example, the frame rate analyzer 30 or portions thereof can be integrated with the controller 12. The reconstruction processor 16 can be integrated with the controller 12. In some embodiments, all three components 12, 16, 30 are contemplated to be integrated as a single unit, for example embodied as a single computer or plurality of interconnected computers executing suitable software to control the scanner 10, perform image reconstruction, and perform frame rate analysis.

Moreover, methods disclosed herein such as the method 50 of FIGURE 4 and the various dynamic magnetic resonance imaging methods disclosed herein with reference to FIGURES 1-4, can be embodied in various ways. These methods may, for example, be embodied as digital instructions stored on a storage medium or media such as a magnetic disk, magnetic tape, optical disk, FLASH memory, random access memory (RAM), read-only memory (ROM), Internet server, or so forth, or by any of various combinations thereof, such instructions being executable by one or more digital processors, controllers, or computers in conjunction with a magnetic resonance scanner to perform one or more of the disclosed methods or variants thereof.

The invention has been described with reference to the preferred embodiments. Modifications and alterations may occur to others upon reading and understanding the preceding detailed description. It is intended that the invention be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

CLAIMSHaving thus described the preferred embodiments, the invention is now claimed to be:

1. An imaging method comprising: acquiring a succession of magnetic resonance images of a subject at a selected frame rate; during the acquiring, measuring a plurality of one-dimensional projections passing through the center of k-space; determining a maximum temporal frequency from the measured one-dimensional projections; and analyzing the selected frame rate based on the maximum temporal frequency.

2. The imaging method as set forth in claim 1, wherein the analyzing includes adjusting the selected frame rate based on the maximum temporal frequency, the method further comprising: adjusting at least one image quality-related parameter based on the adjusted frame rate.

3. The imaging method as set forth in claim 2, wherein the at least one image quality-related parameter includes an image quality-related parameter selected form the group consisting of (i) an image resolution, (ii) a k-space sampling density, and (iii) signal- to-noise ratio (SNR).

4. The imaging method as set forth in claim 1, wherein the analyzing includes adjusting the selected frame rate based on the maximum temporal frequency, the method further comprising: adjusting at least one safety-related parameter based on the adjusted frame rate.

5. The imaging method as set forth in claim 4, wherein the at least one safety-related parameter includes a safety-related parameter selected from the group consisting of (i) SAR and (ii) applied radio frequency energy.

6. The imaging method as set forth in claim 1, wherein the analyzing comprises: setting the selected frame rate to about twice the maximum temporal frequency.

7. The imaging method as set forth in claim 1, wherein the analyzing comprises: setting the selected frame rate to twice the maximum temporal frequency.

8. The imaging method as set forth in claim 1, wherein the acquiring comprises: applying steady-state maintenance radio frequency pulses to maintain a steady state magnetic resonance.

9. The imaging method as set forth in claim 1, wherein the measuring comprises: measuring less than fifteen one-dimensional projections passing through the center of k-space during a time gap between image acquisitions.

10. The imaging method as set forth in claim 1, wherein the measuring comprises: measuring less than ten one-dimensional projections passing through the center of k-space during a time gap between image acquisitions.

11. The imaging method as set forth in claim 1, wherein the measuring comprises: measuring at least one projection at a time immediately adjacent one of the magnetic resonance image acquisitions, the at least one overlapping projection contributing to the immediately adjacent bordering image.

12. The imaging method as set forth in claim 1, wherein the determining comprises: computing a temporal frequency power spectrum from the measured one-dimensional projections; and selecting the maximum temporal frequency as a highest temporal frequency of the temporal frequency power spectrum.

13. The imaging method as set forth in claim 1, wherein the analyzing comprises: providing a perceptible warning conditional upon the maximum temporal frequency exceeding a threshold corresponding to the selected frame rate.

14. The imaging method as set forth in claim 13, wherein the threshold is about one-half of the selected frame rate.

15. The imaging method as set forth in claim 1, wherein the determining performed during a plurality of time gaps in the acquiring of the succession of magnetic resonance images, the projections measured in each time gap being used to determine a maximum temporal frequency at that time gap such that the maximum temporal frequency is determined in approximately real time during the acquiring.

16. A storage medium encoded with instructions executable to control a magnetic resonance scanner (10, 16) to perform the method of claim 1.

17. A magnetic resonance system including means (10, 12, 16, 30, 42) for performing the method of claim 1.

18. An imaging apparatus comprising: a magnetic resonance scanner (10, 16) configured to acquire a succession of magnetic resonance images of a subject at a frame rate and to measure one-dimensional projections passing through the center of k-space during the acquisition; a frame rate analyzer (30) configured to determine a maximum temporal frequency from the measured one-dimensional projections; and a controller (12) configured to adjust the frame rate based on the determined maximum temporal frequency.

19. The imaging apparatus as set forth in claim 18, wherein the controller (12) sets the frame rate to a value of about twice the maximum temporal frequency.

20. The imaging apparatus as set forth in claim 18, wherein the controller (12) is configured to adjust the frame rate in approximately real time based on the maximum temporal frequency determined in approximately real time by the frame rate analyzer.

21. The imaging apparatus as set forth in claim 18, further comprising: a frame rate alarm (42) configured to provide a perceptible warning if the determined maximum temporal frequency exceeds a threshold corresponding to the frame rate.

22. The imaging apparatus as set forth in claim 18, further comprising: a frame rate alarm (42) configured to provide a perceptible warning if the determined maximum temporal frequency exceeds about one-half of the frame rate.

23. The imaging apparatus as set forth in claim 18, wherein the frame rate analyzer includes: a correlator (34) configured to compute a temporal frequency power spectrum based on the measured projections, the maximum temporal frequency being determined from the temporal frequency power spectrum.

24. The imaging apparatus as set forth in claim 23, wherein the frame rate analyzer further includes: a Fourier transform processor (32) configured to compute Fourier transforms of the measured projections prior to input to the correlator (34).

25. An imaging method comprising: acquiring a succession of magnetic resonance images of a subject at a selected frame rate; during the acquiring, measuring a plurality of one-dimensional projections passing through the center of k-space; and determining a minimum frame rate equal to about twice a maximum temporal frequency of the plurality of one-dimensional projections.

26. The imaging method as set forth in claim 25, further comprising: repeating the measuring at a plurality of time gaps between image acquisitions and repeating the determining for each repetition of the measuring to generate a minimum frame rate approximately in real time; and adjusting the selected frame rate in approximately real time during the acquiring to maintain the selected frame rate at about the minimum frame rate.

27. The imaging method as set forth in claim 25, further comprising: providing a perceptible alarm if the minimum frame rate exceeds the selected frame rate.

28. The imaging method as set forth in claim 25, further comprising: spatially filtering the one-dimensional projections before the determining.