Search

The 3D FT variant of spin-echo imaging has previously been successfully used to yield images at microscopic resolution. In obtaining such high resolution, optimization of signal to noise for a given acquisition time is crucial. Using a limited-angle (

Compton scatter of x-ray quanta is the dominant attenuation mechanism in medical diagnostic imaging. In this paper, it is shown that the scatter-to-primary ratio determines the nature and intensity of scatter artifacts in computed tomography (CT) reconstructions and that this ratio, while significantly lower than in radiographic and fluoroscopic examinations, can still be significant in CT. It is found that high spatial-frequency artifacts can arise even though the detected scatter intensity has little or no high-frequency modulation. Reconstructions from x-ray data are presented which show "cupping" as well as dark streaks connecting high-attenuation regions. Correction of the data from measurements of scatter eliminates these artifacts. It is, moreover, observed that the intensity of the artifacts is often diminished when a beam-shaping attenuator is employed. Calculations of scatter intensity are developed from a model which includes single-event and two-event scatter. This analysis is in good agreement with measurements on round water phantoms. Extension to other detector geometries shows, not unexpectedly, that detectors with poorer collimation yield larger scatter artifacts.

A comparison of experimental imaging results obtained with linearly polarized and circularly polarized radiofrequency excitation and reception is presented. Simulation images in good agreement with the experimental scans are described. The simulations are calculated with a model in which a homogeneous, isotropic cylinder of lossy dielectric material and infinite axial extent is immersed in a uniform rf magnetic field perpendicular to the axis. It is found that with the usual linear polarization, reconstructions of uniform objects have regions of decreased intensity. These artifacts are shown to arise from dielectric standing wave effects and eddy currents. The effects become more severe as the frequency or object size is increased, and depend upon the complex conductivity of the object. Results indicate that a significant reduction in the artifact intensity is achieved when circular polarization is employed for both transmission and reception. The expected benefits of circular polarization over linear polarization in reduction of excitation power (up to 50% reduction) and signal-to-noise advantage (√2) have been realized in practice with cylindrical objects and human subjects.

In Nuclear Magnetic Resonance (NMR) the image pixel value is governed by at least three major intrinsic parameters: the spin density N (H), the spin-lattice relaxation time T1, and the spin-spin relaxation time T2. The extent to which the signal is weighted toward one or several parameters is related to the history of the spin system preceding detection. On the simplifying, though not generally warranted assumption that the spin density does not vary significantly in soft tissues, relative tissue contrast can be predicted quantitatively provided the relaxation times are known. Signal intensities and contrast were computed on the basis of the Bloch equations and experimentally determined relaxation times as a function of pulse timing parameters and the data compared with those in images recorded at 0.5T field strength. Significant deviations from the equal density hypothesis were found for gray and white substance. Notably partial saturation but also spin echo and inversion-recovery images are not in full accordance with predictions made on the basis of relaxation times alone.

A procedure is introduced for refabricating any spatially selective excitation pulse to reduce its SAR while preserving its duration and slice profile. Called variable-rate selective excitation, the procedure allows for a variable trade-off of RF amplitude for duration at each sample of the pulse. SAR reduction of 50% is possible with only a mild smearing of the off resonance slice profile. Experimental slice profiles verify the principle.

A new approach to ultrasonic breast imaging is reported. Two-dimensional velocity distributions in tomographic slices transaxial to the breast are reconstructed from transmission time-of-flight projections. The data are displayed both as video images and in numeric format. Results of a clinical investigation in which both malignant and benign lesions have been observed are presented. The possibility of differential diagnosis of pathologies on the basis of velocity is discussed.

We present a method of multislice magnetic resonance imaging that utilizes simultaneous binary-encoded excitation. Signals are acquired from all slices at once, and the images are separated in the reconstruction process. This simultaneous multislice acquisition method has been implemented for multislice spin-echo imaging, and the results are compared with those for a standard interleaved multislice method. Advantages include improved signal-to-noise ratios and flexible slice placement. Phantom and volunteer studies are presented and evaluated in comparison with competing methods.

A method of chemical-shift imaging is described using the invariance of chemical shifts to changes in magnetic field gradients used for frequency encoding of position in imaging. This enables separation of the effects on the observed signal of chemical shift from the effects of different positions along the imaging gradient when the signal is observed with different gradients. A simple implementation for a two-line spectrum is presented using signals observed with normal and reversed imaging gradients. This is used to create “fat” and “water” images of the thigh.© 1985 Academic Press,Inc.

In computed tomography, each projection, or view, is broken up into n data samples. If the interval between samples is d there is an associated Nyquist frequency KN = 1/(2d), which represents the highest spatial frequency retrievable after the sampling process. If higher frequencies are pres

Six patients with pituitary abnormalities and three normal volunteers were evaluated by high field superconductive (1.0, 1.4, or 1.5 Tesla) magnetic resonance (MR) imaging, low field resistive (0.12 Tesla) MR imaging, and contrast-enhanced, high-resolution CT. Four macroadenomas, one microadenoma, and one empty sella were demonstrated. Their morphology and anatomic relationship to the visual pathway and the internal carotid and anterior cerebral arteries were best demonstrated by high field MR imaging. The low field resistive MR studies were least effective in showing the lesions.

The fundamental limit for NMR imaging is set by an intrinsic signal-to-noise ratio (SNR) for a particular combination of rf antenna and imaging subjects. The intrinsic SNR is the signal from a small volume of material in the sample competing with electrical noise from thermally generated, random noise currents in the sample. The intrinsic SNR has been measured for a number of antenna-body section combinations at several different values of the static magnetic field and is proportional to B0. We have applied the intrinsic and system SNR to predict image SNR and have found satisfactory agreement with measurements on images. The relationship between SNR and pixel size is quite different in NMR than it is with imaging modalities using ionizing radiation, and indicates that the initial choice of pixel size is crucial in NMR. The analog of "contrast-detail-dose" plots for ionizing radiation imaging modalities is the "contrast-detail-time" plot in NMR, which should prove useful in choosing a suitable pixel array to visualize a particular anatomical detail for a given NMR receiving antenna.

We present a method for rapid measurement of T1 relaxation times using gradient refocused images at limited flip angles and short repetition times. This "variable nutation" techniques was investigated using a T1 phantom. There was a high correlation between measurements obtained with the variable nutation and partial saturation techniques. The ability of this method to create calculated T1 images is also demonstrated. We conclude that the variable nutation method may allow measurement of T1 relaxation times with a significant reduction in acquisition time compared to partial saturation techniques.

The methodology of sodium-23 (Na-23) imaging is reported in relationship to the physiological factors that determine the chemical environment of the Na-23 nucleus. Contrast resolution is given as a function of imaging time and spatial resolution. Data showing the optimal relaxation time for sodium imaging are given, and the linear quantitative relationship between sodium concentration and voxel intensity for our imaging system is confirmed. The major problem facing in vivo sodium imaging is the ability to differentiate intracellular sodium from extracellular sodium. The sodium in blood serum (extracellular) and packed red blood cells (intracellular) both exhibit biexponential T2 decay. These results indicate that T2 measurements alone will be insufficient for discriminating extracellular from intracellular sodium. Instead, other methods based on the underlying physiological properties of in vivo sodium imaging, such as the diffusion coefficient, will be necessary to truly separate extracellular from intracellular sodium.

Analysis of textural patterns in gray-scale ultrasonography is frequently the basis for clinical diagnosis. It has been found, however, that texture is shaped not only by tissue microstructure but by characteristics of the imaging system as well. This paper describes computer and analytic models which give insight into the role of the imager in texture formation. It is shown that spatial resolution and texture granularity are not simply related, and that axial and lateral texture are determined by unrelated phenomena. Experimental results are given which confirm the modeling.

MR images can be obtained with a 2-sec scan time when an extremely short repetition rate (22 msec), limited flip angle (30 degrees), and gradient refocused echoes are used. Comparison of 415 such images obtained in 29 patients with routine T1-weighted (TR 500, TE 25) and T2-weighted (TR 2000, TE 80) images showed that images free of respiratory artifacts could be obtained in all patients. Although abdominal organs were well seen with 2-sec scan time, overall evaluation of these organs was better on routine T1-weighted images. Vascular structures, however, were seen as well or better on the 2-sec images in 60% of cases. The images were extremely sensitive to field nonhomogeneity, and metallic artifact was exaggerated in five patients with surgical clips. Two-sec MR images provide a rapid method of localizing abdominal organs for further evaluation. The sensitivity to blood flow may assist in the assessment of vascular patency.

A three-dimensional examination of blood vessels is provided using MR data from seven cases. The vascular surfaces are constructed with an algorithm that automatically follows the selected artery or vein and generates a projected three-dimensional gradient shaded image. Fast 3DFT pulse sequences were optimized to enhance the time-of-flight contrast of the intravascular region. By increasing the surface threshold value in a three-dimensional head study, the flesh of a patient's face was peeled away to demonstrate the superfacial temporal artery. Gated cardiac images show the great vessels and cardiac chambers. A three-dimensional view of the aorta shows an irregular surface in the vicinity of an adrenal tumor. 3D MR exams provide a non-invasive technique for assessing vascular morphology in a clinical setting.

This paper describes an "electrodeless" technique for measuring microwave conductivity and dielectric constant of a bulk semiconductor boule without the necessity of fabricating samples. All that is required is that the boule be sufficiently large and that it have one planar surface. The planar surface is butted up to a flange terminating a circular waveguide. The complex TEo01 (circular E-field) mode-reflection coefficient at the surface is then measured with a reflection-coefficient bridge. Because of the particular properties of this mode, no contact impedance occurs as does in the case of the TE10? mode. Thus the measurements yield the true bulk properties of the material. To demonstrate the technique, 48- and 72-GHz measurements of the conductivity and permittivity of a number of semiconductor boules at room temperature are presented and are shown to agree favorably with results obtained by other means.

Techniques employing rectangular waveguides to measure complex permittivity generally contain the tacit assumption that the material fills the waveguide uniformly. With semiconductors this assumption may be poorly satisfied near the waveguide walls due to (1) air gaps, (2) oxide layers, (3) potential barriers, and (4) distributed ``screening'' charges. The uniform‐filling assumption is removed in the present paper by treating the contact region with perturbation theory. The results are found to be simply represented by a lumped‐parameter equivalent circuit. Measurements performed on intrinsic germanium at 10.5 GHz agree well with the theory and indicate that unintentional air gaps and oxide layers dominate the measurements at high conductivities, even for ``tightly'' fitting samples.

This paper describes simple reflection coefficient standards for use with the circular TE/sub 01/ mode. They consist of circular irises inserted directly into the flange connection of a "matched" termination. The reflection coefficient of an iris is formulated exactly and then evaluated to any desired degree of approximation with a digital computer. The results of a 30-mode approximation are presented and shown to agree with 24 GHz measurements of the modulus and phase of /spl rho/ of 7 irises to within about 0.008 relative units and 0.02 radians, respectively.

It is shown in this letter that the measured velocity‐electric field (v‐E) characteristic of n‐type GaAs differes substantially from the intrinsic v‐E characteristic if the doping density of the sample is nonuniform. A simple analysis of the effect is given which compares favorably with a more complete calculation and with experimental results. The heretofore unexplained large variation in measured v‐E characteristics is readily predicted by the present theory if reasonable assumptions are made regarding sample inhomogeneity.

Microwave conductivity of semiconductors determined from transmission or reflection of the dominant mode of rectangular waveguide may contain a significant error caused by the impedance of the contact between the sample and the broad waveguide wall. This error increases with increasing bulk conductivity. The present paper describes a technique which eliminates this source of error by utilizing the TE01° mode of circular waveguide in a ``reflection coefficient bridge.'' Measurements of intrinsic germanium at 48 GHz show no contact influence over the conductivity range between 2 and 2×103 mho/m. In contrast, the conductivity measured with a conventional rectangular waveguide transmission bridge saturates at about 50 mho/m.

Cine magnetic resonance (MR) imaging is a new technique that combines short repetition times, limited flip angles, gradient refocused echoes, and cardiac gating. This technique has a temporal resolution of up to 32 time frames per cardiac cycle and accentuates signal from flowing blood. Cine MR images of 56 valves in 27 patients were evaluated and compared with either Doppler echocardiograms or cardiac catheterization images. An area of decreased signal that correlated spatially and temporally with regurgitant blood flow was seen in all instances in which valvular incompetence was demonstrated on either Doppler echocardiograms or cardiac catheterization images (20 valves). This abnormality was seen in nine of 36 cases without valvular incompetence. Cine MR imaging may be sensitive to turbulence and thus sensitive to valvular regurgitation.

Cine MR imaging provides tomographic images of the heart with both high spatial and high temporal resolution. As many as 32 images per cardiac cycle can be acquired with up to four separate anatomic slices and a total imaging time of 128 cardiac cycles. End-diastolic and end-systolic volumes were determined in 11 patients, and ejection fractions were calculated. The results correlated linearly with those from cardiac catheterization (correlation coefficient of .88). We conclude that cine MR imaging can be used to obtain quantitative information about the heart and has the potential to become a valuable noninvasive means of cardiac evaluation.

A simple analytic model is presented which is used to calculate the impulse response of a computed tomography system for parametric variations of the geometry. The model is useful for fan beam as well as parallel beam geometries. Results show that wide-aperture detector systems have intrinsic resolution limitations that are not mitigated by oversampling. Furthermore, the optimum focal spot size of the X-ray source is found to be dependent on the detector geometry.

Summary and ConclusionIn the high signal -to -noise limit image contrast can be defined as the difference insignal intensity, relative the maximum achievable signal strength. In terms of thisdefinition, contrast can conveniently be predicted on the basis of computed contrast curves,taking into account the tissue relaxation times and pulse timing parameters.In partial saturation images the contrast between tissues of differing T relaxationtime shows a distinct broad maximum. The inversion -recovery sequence, on the other hand,exhibits a more complex behavior in that image contrast reverses sign twice as the inter -pulse interval T is increased. For each pair of adjoining tissue types, which havedifferent T1 relaxation times, two discrete values of T exist for which contrast vanishes.A similar behavior is also found for the spin echo sequence for which a single signreversal occurs. These results are in excellent agreement with signal contrast observedin images recorded under various conditions as far interpulse spacing and cycle repetitiontime are concerned.The present work has been undertaken with the objective to quantify image contrast andultimately to develop pathology- specific pulse protocols. The latter may be achieved byestablishing a data bank of relaxation parameters in normal and pathologic tissue andby optimizing the pulse timing parameters in the manner described.AcknowledgementsThe authors are indebted to Nancy Grigsby for her help with the experimental work andthe scanning of volunteers.References1. Shaw, D., Fourier Transform NMR Spectroscopy, Elsevier, 1976, Chap. 2.2. Kaufman, L., Crooks, L.E., and Margulis, A.R., Nuclear Magnetic Resonance Imaging inMedicine, Igaku -Shoin , 1982, Chap. 6.3. Canet, D., Levy, G.C., and Peat, I.R., J. Mag. Res. 18, 199 (1975).4. Edelstein, W.A., Bottomley, P.A., Hart, H.R., Leue, W.M., Schenck, J.F., and Redington,R.W., in NMR Imaging, Proceedings of the Bowman Gray Symposium, Winston -Salem, N.C.

Coronal nuclear magnetic resonance (NMR) abdominal imaging was performed on a normal volunteer. The scan technique and anatomic features are described. Coronal abdominal scanning optimizes vascular imaging in the abdomen and is helpful in defining the anatomic relationships of a variety of structures.

Conventional chemical shift magnetic resonance (MR) imaging with the phase contrast technique has a number of limitations with respect to quantitative accuracy. The hypothesis of this study is that the accuracy of phase contrast chemical shift MR may be improved by increasing the number of basis images from two to three. Water and fat images were obtained from phantoms and volunteers with a 1.5 T MR system using double and triple acquisition phase contrast chemical shift methods. Longitudinal relaxation time and relative water and fat content were calculated from these basis images. The T1 relaxation times of the aqueous component of composite phantoms were determined more accurately using the triple acquisition method than with the double acquisition method. In vivo studies demonstrate that the triple acquisition method separated fat and water signals more accurately and showed less field inhomogeneity dependence than the modified double acquisition method. The new method also provided a map of static field magnetic inhomogeneity and tissue magnetic susceptibility. The triple acquisition phase contrast chemical shift imaging technique should improve the prospect for quantitative tissue characterization in clinical MR.

Dark streaks connecting the petrous bones are often observed in cranial transverse section CT reconstructions. These artifacts are usually only slightly diminished by two-pass beam hardening corrections. However, it is found that by narrowing the slice thickness the artifacts are substantially reduced. In this paper, it is shown that axial partial volume effects can account for the presence of the artifacts. These axial partial volume effects occur when, at any point in the slice, the object has axial variations in attenuation. In such cases the logarithm of the integrated intensity measured by the detector is not a linear function of the integrated attenuation (even for monochromatic beams). This nonlinearity causes inconsistencies in the data set which in turn can cause streaks in the image. We have studied the partial volume effect using computer simulation. Algorithms are presented whose purpose is to correct for these effects by estimating the axial variation using neighboring slices. These correction algorithms are successful in computer simulation cases but failed with clinical data. It is concluded that no practical correction method is viable unless overlap scanning is employed. However, thin-slice scanning for sections where these artifacts are common is perhaps a more preferable solution.

The medical community has long practiced the art of nondestructive evaluation in its various diagnostic imaging branches. Only within about the last decade, however, has there been an attempt to provide quantitative information to the clinical radiologist. The most important example of a system providing such information is the computed x-ray tomography (CT) scanner. The resulting digitally created images portray maps of an approximation to the electron density at discrete sampling points within the subject. Density resolution of the order of 0.1% and spatial resolution reaching 0.5 mm have become readily available.

Low field strength (0.12 Tesla resistive) and high field strength (1.0, 1.4, or 1.5 Tesla superconductive) magnetic resonance imagers were compared for their ability to detect central nervous system lesions. Sixteen adult patients with known lesions and three normal volunteers were studied. Contrast-enhanced computed tomography was used as the standard. The data demonstrate that imaging at high magnetic field strength is superior to low field strength imaging for the detection and delineation of lesions. This finding can be explained by the superior signal-to-noise ratio achievable at the higher magnetic field strengths. High field MR imaging was also found to outperform CT in demonstrating anatomic details and relationships. It is predicted that the use of low saturation (e.g., long TR spin echo technique) will make the gain in contrast-to-noise ratio even more significant.

Respiration effects and cardiac pulsatility can induce signal modulations in functional MR image time series that increase noise and degrade the statistical significance of activation signals. A simple image-based correction method is described that does not have the limitations of k-space methods that preclude high spatial frequency correction. Low-order Fourier series are fit to the image data based on time of each image acquisition relative to the phase of the cardiac and respiratory cycles, monitored using a photoplethysmograph and pneumatic belt, respectively. The RETROICOR method is demonstrated using resting-state experiments on three subjects and compared with the k-space method. The method is found to perform well for both respiration- and cardiac-induced noise without imposing spatial filtering on the correction. Magn Reson Med 44:162‐167, 2000. © 2000 Wiley-Liss, Inc.

Functional magnetic resonance imaging (fMRI) based on multiple-echo T(2)* mapping has attracted much attention recently. The contrasts in the parametric T(2)* maps are usually too low to allow direct image registration. In this study, an image registration strategy has been proposed for single-shot multi-echo data sets acquired for dynamic T(2)* mapping. We performed image registration of the T(2)*-weighted images before the calculation of the T(2)* parameter maps using two different strategies. One is to perform separate image registration on each echo and the other is to use the same motion correction parameters extracted from the second echo for all the data. Both strategies increase the number of activated voxels and reduce the effective noise level. The results also indicate that, for a single-shot dual-echo image data set, it is slightly preferable to use the second echo for direct image registration and then apply the same motion correction parameters to the first echo images. J. Magn. Reson Imaging 1999;10:154-158.

In this study, the hemodynamic response and changes in oxidative metabolism during functional activation were measured using three functional magnetic resonance imaging (fMRI) techniques: the blood oxygenation level-dependent (BOLD) technique, flow-sensitive alternating inversion recovery (FAIR), and bolus tracking (BT) of an MR contrast agent. With these three techniques we independently determined changes in BOLD signal, relative cerebral blood flow (rCBF), and cerebral blood volume (rCBV) associated with brain activation in eight healthy volunteers. In the motor cortex, the BOLD signal increased by 1.8 +/- 0.5%, rCBF by 36.3 +/- 8.2% (FAIR), and 35.1 +/- 8.6% (BT), and rCBV by 19.4 +/- 4.1% (BT) in response to simultaneous bilateral finger tapping. In the visual cortex, BOLD signal increased by 2.6 +/- 0.5%, rCBF by 38.5% +/- 7.6 (FAIR), and 36.9 +/- 8.8% (BT), and rCBV by 18.8 +/- 2.8% (BT) during flickering checkerboard stimulation. Comparing the experimentally measured rCBV with the calculated rCBV using Grubb's power-law relation, we conclude that the use of power-law relationship results in systematic underestimate of rCBV.

In an NMR imaging method employing multiple spin-echo pulse sequences, a method is disclosed for eliminating ghosting artifacts in an NMR image. The method comprising the steps of (1) maintaining a static magnetic field along a first axis of an NMR imaging sample, (2) exciting a plurality of nuclear spins in a region of said NMR sample by irradiating said sample with an RF pulse, (3) irradiating said NMR sample with a series of 180° RF pulses such that each RF pulse inverts the orientation of a first group of the excited nuclear spins by 180°, whereupon inherent imperfections in at least one of the 180° pulses cause a second group of excited nuclear spins to change orientation by an angle other than 180°, (4) applying first and second magnetic field gradient pulses to the NMR sample before and after, respectively, the application of at least one of the 180° RF pulses thereby causing no net accumulation of phase in the first group of excited nuclear spins, (5) applying an imaging gradient after the 180° RF pulse in order to focus the plurality of nuclear spins and thereby create an NMR spin-echo signal, (6) sampling the NMR signal during a predetermined time interval between 180° pulses in order to detect the spin-echo signal; and (7) modulating each pair of first and second magnetic field gradients with respect to other pairs associated with the series of 180° RF pulses in order to prevent refocusing of the second group of nuclear spins for the duration of all of the time intervals during which the NMR signal from the sample is detected for the purpose of constructing an image.

This paper reports a direct determination of the relative, static permittivity of the atomic lattice of InSb at 77°K. The measurements were performed on relatively pure (ND≤1014 cm−3) n‐type single‐crystal samples at 70.1 GHz by utilizing a strong magnetic field (∼40 kG) to immobilize the charge carriers in a direction parallel to the applied TE01°‐mode electric field. In this manner, the rf losses normally associated with ``drift'' motion of the charge carriers were made negligible, thus permitting the lattice component of the permittivity to be accurately resolved. The measured relative value, es=17.78±0.15, compares favorably with several values previously calculated from measurements of the ir reststrahlen spectrum.

The relative dielectric constant er of high‐resistivity GaAs has been measured at 70.243 GHz as a function of temperature between 100 and 300°K. The measuring technique utilized a circular E field (TE°01) mode reflection‐coefficient bridge. Estimated relative and absolute accuracies of the measurements are ±0.2% and ±0.5%, respectively. The results are found to fit the equation er(T) = er(0){1 + αT} where er(0) = 12.73 ±.07 and α = (1.2 ± 0.1) × 10−4. At room temperature (295°K) the relative permittivity is er = 13.18 ±.07.

A technique has been developed which reduces blood-flow artifacts in multi-echo, multi-slice magnetic resonance (MR) imaging by rephasing spins which have arbitrary velocity as well as static spins. The technique requires tailoring of the gradient structure along at least one of the three axes so as to null out in the preferred embodiment the zeroth and first moments of the gradient distributions.

A method is provided for enhancing image signal-to-noise ratio (SNR) by combining NMR acquired images of differing pulse sequence timing. In one embodiment, a plurality of multi-echo images acquired at different spin-echo times are used to compute values of spin-spin relaxation (T 2 ) and pseudospin density (PD) for each pixel in the multi-echo images. These values are used to generate new images at a spin-echo time TE c for which actual measurements may not have been taken. In another embodiment, the calculated T 2 images are used to extrapolate (by multiplying each pixel by an appropriate factor) each of the acquired images to some spin-echo time TE c . The extrapolated image can now be averaged to produce an image with improved SNR. In yet another embodiment, similar techniques are used to improve the SNR of images acquired by varying parameters, such as pulse sequence repetition time TR, which effect the spin-lattice (T 1 ) dependence of the image.

In a system for reducing artifacts in a desired image due to substantially periodic variations in the imaging signal detected from a portion of an object under examination, a method and apparatus for generating from a signal y(t) related to the variations and whose values are not equally likely, a signal φ(t) whose values are substantially equally likely for use in distributing events substantially evenly over the variations. In order to provide the signal φ(t), an apparatus is disclosed for executing the following steps: (a) establishing a historical data base derived from values of the signal y(t) for a period of time preceding the present time, (b) measuring parameters related to the present value of the signal y(t); and (c) producing a value for φ(t) using the present measured parameters and the historical data base; wherein the values of φ(t) are more evenly distributed than the values of y(t).

A method for reducing image artifacts due to signal variations in the course of examining a subject using nuclear magnetic resonance (NMR) techniques includes the acquisition of scan data for imaging the object. The scan data is composed of a number of views. The acquisition of each view includes the implementation of a pulse sequence to generate an NMR signal and application of a magnetic gradient along at least one dimensional axis of the object. The magnetic field gradient is characterized by a parameter (e.g., amplitude or direction) adjustable from view to view to encode spatial information into the NMR signal. The parameter value is selected just prior to implementation of the pulse sequence for that view and depends on the phase of the signal variation at that point in time. The final view order depends on the measured phase during the scan. In this manner, the view order can be continuously adjusted to obtain optimum artifact reduction. The method is applicable to various NMR imaging techniques, including Fourier transform and multiple angle projection reconstruction.