WO1989012834A1

WO1989012834A1 - Spectroscopy methods

Info

Publication number: WO1989012834A1
Application number: PCT/US1989/002642
Authority: WO
Inventors: Kamil Ugurbil; Michael Garwood
Original assignee: Regents Of The University Of Minnesota
Priority date: 1988-06-21
Filing date: 1989-06-16
Publication date: 1989-12-28
Also published as: EP0420915A4; EP0420915A1; JPH03505292A

Abstract

Several magnetic resonance imaging (MRI) methods using adiabatic excitation are disclosed. One method accomplishes slice selection with gradient modulated adiabatic excitation. Another method employs slice selection with adiabatic excitation despite large variations in B1 magnitude. There is also described 1H spectroscopy using solvent suppressive adiabatic pulses.

Description

SPECTROSCOPY METHODS

Technical Field of the Invention The present invention relates generally to spectroscopy, and more particularly to methods for slice selection and solvent suppression using adiabatic excitation.

Background of the Invention Magnetic resonance imaging (MRI ) is now an important imaging technique in medicine. There are described herein several MRI methods using adiabatic excitation. One method accomplishes slice selection with gradient modulated adiabatic excitation. Another method employs slice selection with adiabatic excitation despite large variations in B₁ magnitude. There is also described ¹H spectroscopy using solvent suppressive adiabatic pulses. The methods described herein relate to the adiabatic pulses and methods described in U.S. Patent Application Serial No. 032,059, entitled "Amplitude and Frequency/Phase Modulated Pulses to Achieve Plane Rotations of Nuclear, Spin Magnetization Vectors, with Inhomogeneous B₁ Fields", the entire disclosure of which is hereby incorporated by reference herein.

Brief Description of the Drawings Fig. 1 is a plot of B^e(t) generated by a simple version of GMAX consisting of tanh modulated B₁ amplitude, and sech modulated pulse frequency and gradient magnitude;

Fig. 2 is a plot of a computer calculated slice profile obtained with the tanh/sech version of GMAX;

Fig. 3A is a plot of signal-to-noise obtained with Solvent Suppressive Adiabatic Pulses;

Fig. 3B is a plot of signal-to-noise obtained with binomial pulses. Detailed Description of the Invention

SLICE SELECTION WITH GRADIENT MODULATED ADIABATIC EXCITATION

Because of their extreme tolerance to RF inhomogeneity, frequency selective adiabatic inversion pulses are becoming increasingly popular for localized spectroscopic and imaging studies carried out with surface coils. They offer superior slice definition and do not suffer from sensitivity loss caused by non-uniform spin excitation, a problem inherent with conventional pulses. With frequency selective B₁ insensitive inversion pulses, slice selection is achieved by applying the pulse in the presence of a constant B₀ gradient field while the pulse amplitude and frequency (or phase) are modulated. We describe here a new gradient modulated adiabatic excitation pulse (GMAX) which, in addition to amplitude and frequency modulation, employs time-dependent gradient modulation in order to achieve slice selective excitation.

To clearly illustrate the principles of GMAX, consider a frame of reference (axes x', y', z') precessing at the instantaneous frequency of the pulse. Under ideal adiabatic conditions, motion of the initial longitudinal magnetization will parallel the trajectory of the effective field B^e(t), which is defined in this frame by the vectorial sum of B_1(t) and Δω(t). B₁(t) is RF amplitude in rads/sec and Δω(t) equals the difference between the instantaneous pulse and spin frequencies (rads/sec). Fig. 1 shows B^e(t) generated by a simple version of GMAX consisting of tanh modulated B₁ amplitude, and sech modulated pulse frequency and gradient magnitude. The terms A, v, T, and B are, respectively, pulse frequency modulation amplitude (in Hz), a unitless term accounting for spatial variation in B₁ magnitude, pulse duration, and the limit of modulation. The quantity g(r) in the expression for Δω(t) represents the maximum B₀ gradient strength (rads/sec) as a function of position, r.

Unlike adiabatic inversion (Silver, M.S., R.I. Joseph, and D.I. Hoult, J. Magn. Reson. 59,347,1984) or excitation (Johnson, A.J., K. Ugurbil, and M. Garwood. Abstract submitted for 7th Annual Society of Magnetic Resonance in Medicine meeting, 1988) pulses where the field gradient is time-invariant, GMAX creates a gradient-dependent node in Δω(t) by executing equivalent time modulation of both pulse frequency and B₀ gradient strength. This node exists in a plane lying perpendicular, to the gradient direction in a region where the field gradient magnitude exactly equals the pulse frequency amplitude, i.e. when g(r)= 2 π A. Since Δω(t) will be positive on one side of this plane and negative on the other, the final transverse magnetization on opposite sides will differ in phase by 180°. A second implementation of GMAX with reversed gradient modulation yields a mirror image of the first response reflected about a point where the gradient modulation is zero. Addition of these two signals defines a slice of excitation of width 4 π A. Slice position and width can be varied by manipulating A and/or gradient magnitude. By repeating this sequence with three orthogonal field gradients, 3-dimensional localization can be achieved in a manner analogous to 3-dimensional ISIS (Ordidge, R.J., A. Connelly, and J.A.B. Lohman. J. Magn. Reson. 66, 283, 1986). Fig. 2 shows a computer calculated slice profile obtained with the tanh/sech version of GMAX, B=5.3.

Although the included example of GMAX consists of a tanh/sech modulation pair, other functions can be used provided the boundary conditions remain unchanged; specifically, numerically optimized modulation schemes can be used to enhance the pulse with respect to off-resonance performance and B₁ insensitivity (Ugurbil, K., M. Garwood, and A. Rath J. Magn. Reson. in press, 1988). The Silver, Johnson, Ordidge and Ugurbil references specified above are hereby incorporated by reference herein. SLICE SELECTION WITH ADIABATIC EXCITATION DESPITE LARGE VARIATIONS IN B₁ MAGNITUDE Frequency selective excitation pulses used in NMR imaging and spectroscopy are Bj sensitive and cannot induce uniform excitation, or maintain a uniform slice profile in the presence of large variation in B₁ magnitude. Consequently, in surface coil studies using these pulses, it is virtually impossible to select a distortion free slice that is perpendicular to the plane of the surface coil. The only methods capable of defining uniform slices over large variations in B₁ have relied on adiabatic inversion (Silver, M.S., R.I. Joseph, and D.I. Hoult, J. Magn. Reson. 59, 347, 1984) or excitation (Johnson, A.J., M. Garwood, and K. Ugurbil. Abstract submitted for 7th Annual Society of Magnetic Resonance in Medicine meeting, 1988) pulses that require the addition of two separate acquisitions. These pulses are unsuitable for multislice imaging or spectroscopic and imaging studies in the presence of motion. Here, we report a new adiabatic excitation pulse that is both slice selective and B₁ insensitive and does not require multiple acquisitions. Described in terms of amplitude and frequency modulation, this slice selective adiabatic excitation or SSAX pulse consists of two functionally distinct but contiguous segments. The initial segment combines constant B₁ amplitude with a large frequency sweep; the subsequent segment couples decaying B₁ amplitude with zero frequency modulation. Taking B₁ to be along x' in a reference frame which rotates at the instantaneous frequency of the pulse (axes x', y', and z ' ) , the effect of SSAX can be visualized as follows: SSAX first creates a distribution of spin orientations in the x'z' plane as a function of frequency offset so that only spins with zero offset point along x'. Subsequently, spins not along x' are taken either back up to z' (positive offset) or down to -z' (negative offset) whereas those along x ' ( zero offset ) are allowed to remain in the transverse plane.

For illustrative purposes, the following example is a simple version of a SSAX pulse constructed with tangent and sech functions.

B₁(t) = 2πAV

Δω(t) = 0

B₁ is RF amplitude in rads/sec, Δω is the difference between the instantaneous pulse and spin Larmor frequencies in rads/sec, A*tan [πq/2] is the frequency modulation amplitude in Hz, v is a unitless parameter equal to the (peak B₁)/2πA ratio, T is pulse duration, and q and β are chosen to set the modulation limits. When B₁ is inhomogeneous, both B₁ and v depend on spatial coordinates.

While the above example is written using tan and sech functions, provided the boundary values remain the same, other modulation schemes can be used; in par ticular, numerically optimized modulation routines can be used to improve off-resonance performance and B₁ insensitivity of the pulse (Ugurbil, K., M. Garwood, and A. Rath. J. Magn. Reson. in press, 1988). Projected, applications include combining SSAX with a single adiabatic spin excitation to achieve solvent suppression while exciting other frequencies, combining two SSAX pulses with a 90° adiabatic rotation from the x'y' plane to the z' axis to define a 2-dimensional column" of excitation in one pulse train. Similarly, three SSAX pulses can be combined with two 90° adiabatic rotations to define in a single pulse train a 3-dimensional volume of excitation which can be used in spectroscopic localization. The slice profile of the current version of this pulse is not square. However, there are no B₁ insensitive and slice selective pulses which, do not require multiple acquisitions. Therefore, this pulse would be preferred in all applications in the presence of inhomogeneous B₁'s where multislice capability is desired, motion is present, and/or subtraction errors are significant. The disclosures of the above-specified Silver, Johnson and Ugurbil references are hereby incorporated by reference herein.

¹H SPECTROSCOPY USING SOLVENT SUPPRESSIVE ADIABATIC PULSES (SSAP)

In vivo ¹H NMR spectroscopy is usually executed with surface coils, which provide enhanced sensitivity for most applications although the B₁ field is extremely inhomogeneous. Pulse sequences designed to selectively suppress the H₂O signal have been based on RF pulses which are sensitive to variations in B₁ magnitude (P.J. Hore. J.Magn. Reson. 55, 383(1983)). When using surface coils to transmit rectangular or amplitude modulated pulses, sample regions inevitably experience nutation angles, Θ that are multiples of 180°, and signals arising in regions where θ=90° may be partially cancelled by signals produced where Θ=270°. In addition, the frequency response of such pulses depends upon Θ, and is therefore a function of spatial coordinates when inhomogeneous RF coils are employed for RF transmission.

Recently, we described amplitude and frequency/phase modulated pulses which are highly insensitive to B₁ inhomogeneity and can achieve uniform 90° (K. Ugurbil, M. Garwood, and M.R. Bendall, J. Magn. Reson. 72, 177(1987)) and 180° (M.R. Bendall, M. Garwood, K. Ugurbil, and D.T. Pegg, Magn. Reson. Med. 4, 498(1987), K. Ugurbil, M. Garwood, A.R. Rath, and M.R. Bendall, J. Magn. Reson., in press) plane rotations across the surface coil active volume provided the adiabatic condition (i.e., | B^e/(dα /dt)| >> 1) is satisfied throughout the pulse. These pulses are composed of segments during which the effective field, B^e, rotates 90° with respect to a frame rotating with the instantaneous frequency of the pulse. The 90° and 180° plane rotation pulses, BIR-2 and BIREF-1, are composed of 4 and 2 such segments, respectively. When a delay period, t, equal to ½v is placed between the first and second 90° segments of BIR-2, spins with precessional frequencies equal to ±nv are returned to the z'-axis at the end of the pulse, where n is an odd integer and v is the frequency offset (in Hz) in the rotating frame. Likewise, spins with frequencies equal to ±nv do not refocus when a delay period equal to ½v is placed in the middle of refocusing pulse BIREF-1. Consequently, pulses BIR-2 and BIREF-1 can be modified to achieve solvent suppression, where the frequency response (cos(2 π v t) and cos( π v t), respectively) and spatial dependence are highly invariant across a wide range of B₁ magnitude.

BIR-2 and BIREF-1 solvent suppressive adiabatic pulses (SSAP) were used for excitation and refocusing in a spin echo sequence to obtain in vivo ¹H spectra of rat brain. Spectra were acquired using a GE CSI-II spectrometer equipped with a 40 cm, 4.7 T magnet. An 8 mm diameter surface coil placed over the rat head was used. t was adjusted to produce a null at the H₂O resonance and yield maximal signal 500 Hz either side of H₂O. Fig. 3A shows the results obtained with SSAP after 48 scans using a repetition time of 2 sec and an echo time (TE) of 136 msec. For comparison. Fig. 3B shows the results obtained from the same animal using a spin echo with binominal pulses 1-3-3-1 and 2-6-6-2 (P.J. Hore J. Magn. Reson. 55, 383(1983)). For this latter experiment, the pulse lengths were adjusted so that the cumulative excitation flip angle equaled 135° on resonance at the coil center. All other parameters remained constant. Because adiabatic pulses can uniformly excite spins throughout the surface coil active volume, the spectral signal-to-noise obtained with SSAP (Fig. 3A) was increased relative to that obtained with the binomial pulses (Fig. 3B). The disclosures of the above-specified Bendall, Hore and two Ugurbil references are hereby incorporated by reference herein.

Although the invention has been described herein in its preferred form those skilled in the art will readily recognize that many modifications and changes may be made thereto without departing from the spirit and scope of the claims appended hereto.

Claims

What is claimed is:

1. Method for slice selection in a sample under the influence of a constant B₀ gradient field comprising the steps of: a) providing a surface coil proximate said sample; b) energizing said surface coil with an adiabatic excitation pulse; and c) modulating the gradient of said pulse over time in order to achieve slice selective excitation.

2. A method of slice selection in a sample under the influence of a constant B₀ gradient field comprising the steps of: a) providing a surface coil proximate said sample; and b) exciting said surface coil to produce an adiabatic excitation pulse comprising two functionally distinct but contiguous segments, and an initial segment of said pulse including a constant B₁ amplitude with a large frequency sweep, the subsequent segment of said pulse coupling a decaying B₁ amplitude with zero frequency modulation.

3. A method for solvent suppression in ¹H spectroscopy in a sample under the influence of a constant B₀ gradient field comprising the steps: a) providing a surface coil proximate said sample; and b) energizing said coil to produce an adiabatic pulse comprising segments separated by a predetermined delay period to achieve solvent suppression.