Method and system for measuring and compensating for eddy currents induced during NMR imaging operations

• US 6,448,773 B1
• Filed: 02/24/2000
• Issued: 09/10/2002
• Est. Priority Date: 02/24/2000
• Status: Expired due to Term

• Current Assignee: Toshiba America MRI Incorporated (Toshiba Corporation)
• Original Assignee: Toshiba America MRI Incorporated (Toshiba Corporation)
• Inventors: Zhang, Weiguo
• Primary Examiner(s): ARANA, LOUIS M
• Application Number:
  US09/512,323
• Time in Patent Office: 929 Days
• Field of Search: 324/300, 324/306, 324/307, 324/309, 324/312, 324/314, 324/318, 324/322
• US Class Current: 324/309
• CPC Class Codes: G01R 33/56518 due to eddy currents, e.g. caused by switching of the gradient magnetic field

Abstract

A method and apparatus is disclosed for measuring and compensating the effects of eddy currents induced during NMR imaging operations. A cubic or cylindrical sample is placed in the imaging volume of a MRI system at a position centrally located with respect to the main magnetic field and oriented with its longitudinal axis parallel to a desired measuring direction. A magnetic field gradient pulse is applied for inducing eddy currents as well as for generating a slice-selective spin-echo signal. The spin-echo signal is acquired immediately after the termination of each eddy-current inducing gradient pulse. Two slices are selected along the desired measurement direction at symmetrical equal distance from the center of the main magnetic field. Two spin-echo signals are acquired for each slice with the polarity of the eddy-current inducing gradient pulse reversed between the two echo signals. Quantitative values for eddy-current induced field gradients and B₀ oscillations are determined based on the precessing frequencies of the acquired NMR signals. NMR imaging is improved by compensating for eddy currents effects by applying the quantified values of the field gradients and B₀ oscillations to set an appropriate pre-emphasis network. Gradient pulses in MRI/MRS pulse sequences may also be selectively pre-distorted or modified to compensate for resulting gradient-switching induced eddy currents. Other aspects of the disclosed method include measuring the time course of gradient switching, altering the pulse sequences to measure eddy currents having long time constants, repeatedly measuring the eddy currents to assist in pre-emphasis adjustments, and measuring EC-induced field gradients and B₀ oscillation in the presence of moderately large background field inhomogeneities.

Claims

1. A method for evaluating eddy currents induced during gradient switching in an MRI system having a static magnetic field, B₀, and gradient coils to induce gradients in the static field, the method comprising the steps of:

a) placing an MR active test object at a center of an imaging volume of the MRI system and aligning the object along a predetermined direction for measurement of eddy currents;

b) applying a first gradient pulse to a gradient coil to produce a switching gradient in the magnetic field B₀ and acquiring signal data samples from a first pair of spin-echo NMR signals corresponding to two MRI slices located symmetrically apart along said measurement direction away from a center of the magnetic field B₀;

c) applying a second gradient pulse to the gradient coil, the second gradient pulse having a polarity opposite to the first gradient pulse, and acquiring signal data samples from a second pair of spin-echo NMR signals corresponding to the slices of step (b);

d) determining precessing frequencies from acquired signal data samples for each spin-echo signal of the first and second pair of spin-echo NMR signals; and

e) quantifying field gradients and/or field oscillations produced due to eddy currents from the precessing frequencies determined in step (d).

2. A method as in claim 1 wherein quantitative values for precessing frequencies of spin-echo signals are determined from acquired NMR signal data samples according to the following relationships:

$f_n<br><br>\left(i\right)=\frac{u<br><br>\left(i\right)<br><br>v<br><br>\left(i\right)-v<br><br>\left(i\right)<br><br>u<br><br>\left(i\right)}{u^2<br><br>\left(i\right)+v^2<br><br>\left(i\right)}$$\mathrm{with}$$u<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)+S_n^u<br><br>\left(i\right)}{2}$$v<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)+S_n^v<br><br>\left(i\right)}{2}$${\hspace{0.17em}<br><br>}_{\stackrel{.}{u}}<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)-S_n^u<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$${\hspace{0.17em}<br><br>}_{\stackrel{.}{v}}<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)-S_n^v<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$

where f_n(i) for n=1, 2, 3, 4 represents precessing frequencies of four NMR signals corresponding to said first and second pair of spin-echo signals;

i for i=0,1, 2, . . . N−

2 represents a particular data sample with N being a predetermined number of data samples acquired from a spin-echo NMR signal;

s_n^u and s_n^v are real and imaginary parts of complex NMR signal data acquired;

{dot over (u)}(i) and {dot over (v)}(i) are corresponding time derivatives of u(i) and v(i), respectively;

Δ

t is a sampling rate of acquired NMR signal data; and

wherein quantitative values for a time-varying field gradient and a time-varying B₀ oscillation are determined according to the following relationships;

$\mathrm{\gamma <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>G<br><br>\left(t\right)=\frac{f_1<br><br>\left(t\right)-f_2<br><br>\left(t\right)-f_3<br><br>\left(t\right)+f_4<br><br>\left(t\right)}{4<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>Z}$$\mathrm{\gamma <br><br>\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>B<br><br>\left(t\right)=\frac{f_1<br><br>\left(t\right)+f_2<br><br>\left(t\right)-f_3<br><br>\left(t\right)-f_4<br><br>\left(t\right)}{4}$

where f_n(t) represents a time-varying precessing frequency of a spin-echo signal;

Δ

G(t) and Δ

B(t) are, respectively, the time-varying field gradient and time-varying B₀ oscillation produced by eddy currents; and

γ

is a gyromagnetic ratio value of an appropriate nuclear species.

3. A method as in claim 1 wherein an applied field gradient used for slice selection also serves as an eddy current inducing gradient pulse for which eddy currents are evaluated and said slice-selected spin-echo NMR signals are generated by applying a selective 90°

RF pulse followed by another selective 180°

pulse during the application of said field gradient.

4. A method as in claim 3 for evaluating short time-constant eddy currents wherein said spin-echo NMR signals are acquired immediately after switching of said gradient pulse and are acquired during a predetermined period of time for which quantitation of EC-induced field gradient and B₀ oscillation is desired.

5. A method as in claim 3 for characterizing a gradient switching time course wherein acquisitions of said spin-echo signals are started before said gradient pulse is switched to a different target value and are continued to cover the entire transition period of the gradient switching, and wherein said spin-echo signals are used to characterize the time course of the gradient switching.

6. A method as in claim 1 for evaluating long time-constant eddy currents wherein an MRI pulse sequence is used which produces NMR signals that are generated after termination of said gradient pulse and NMR signal data samples are repeatedly acquired for a predetermined period of time for which quantitation of eddy currents is desired.

7. A method as in claim 6 wherein said MRI sequence is a slice selective field-echo sequence.

8. A method as in claim 6 wherein said MRI reference is a slice selective spin-echo sequence.

9. A method as in claim 6 wherein said MRI sequence is a slice selective stimulated-echo sequence.

10. A method as in claim 1 wherein MRI system response to adjustments of a pre-emphasis network associated with the MRI system are evaluated, further comprising the steps of:

(i) producing time-varying EC-induced field gradient and B₀ oscillation, Δ

G(t) and Δ

B(t), respectively, after switching off said EC-inducing gradient pulse;

(ii) changing the pre-emphasis network settings;

(iii) repeating steps (i) and (ii) a predetermined number of times; and

comparing time course graphs of EC gradients B₀ oscillation generated in step (i) to evaluate MRI system response to changes in pre-emphasis network settings.

11. A MRI system for evaluating eddy currents induced during magnetic gradient switching wherein an MR active test object is positioned at a center of an imaging volume of the MRI system and aligned along the measurement direction, the MRI system having a static magnetic field, B₀, and gradient coils to induce gradients in the static field, comprising:

means for applying a first gradient pulse to a gradient coil and acquiring a first pair of spin-echo signal samples corresponding to two slices located symmetrically apart along said measurement direction away from a center of the magnetic field B₀; and

means for applying a second gradient pulse to the gradient coil, the second gradient pulse having a polarity opposite to the first gradient pulse, and acquiring a second pair of spin-echo signal samples corresponding to said two slices; and

means for quantitatively determining magnetic field gradients and/or magnetic field oscillations due to eddy currents according to precessing frequencies of acquired spin-echo signals.

12. A MRI system as in claim 11 wherein quantitative values for precessing frequencies of spin-echo signals are determined from acquired NMR signal data samples according to the following relationships:

$f_n<br><br>\left(i\right)=\frac{u<br><br>\left(i\right)<br><br>v<br><br>\left(i\right)-v<br><br>\left(i\right)<br><br>u<br><br>\left(i\right)}{u^2<br><br>\left(i\right)+v^2<br><br>\left(i\right)}$$\mathrm{with}$$u<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)+S_n^u<br><br>\left(i\right)}{2}$$v<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)+S_n^v<br><br>\left(i\right)}{2}$${\hspace{0.17em}<br><br>}_{\stackrel{.}{u}}<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)-S_n^u<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$${\hspace{0.17em}<br><br>}_{\stackrel{.}{v}}<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)-S_n^v<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$

where f_n(i) for n=1, 2, 3, 4 represents precessing frequencies of four NMR signals corresponding to said first and second pair of spin-echo signals;

i for i=0,1, 2, . . . N−

2 represents a particular data sample with N being a predetermined number of data samples acquired from a spin-echo NMR signal;

s_n^u and s_n^v are real and imaginary parts of complex NMR signal data acquired;

{dot over (u)}(i) and {dot over (v)}(i) are corresponding time derivatives of u(i) and v(i), respectively;

Δ

t is a sampling rate of acquired NMR signal data; and

wherein quantitative values for a time-varying field gradient and a time-varying B₀ oscillation are determined according to the following relationships;

$\mathrm{\gamma <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>G<br><br>\left(t\right)=\frac{f_1<br><br>\left(t\right)-f_2<br><br>\left(t\right)-f_3<br><br>\left(t\right)+f_4<br><br>\left(t\right)}{4<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>Z}$$\mathrm{\gamma <br><br>\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>B<br><br>\left(t\right)=\frac{f_1<br><br>\left(t\right)+f_2<br><br>\left(t\right)-f_3<br><br>\left(t\right)-f_4<br><br>\left(t\right)}{4}$

where f_n(t) represents a time-varying precessing frequency of a spin-echo signal;

Δ

G(t) and Δ

B(t) are, respectively, the time-varying field gradient and time-varying B₀ oscillation produced by eddy currents; and

γ

is a gyromagnetic ratio value of a n appropriate nuclear species.

13. In an MRI system having a background static magnetic field, B₀, and gradient coils to induce switched gradients in the static field, a method for compensating eddy current (EC) effects and B₀ oscillations produced during an MRI sequence, comprising the steps of:

a) determining quantified values for an EC gradient and a B₀ oscillation induced by a switched gradient within the MRI system, wherein quantitative values for an EC gradient and a B₀ oscillation are determined according to the following two relationships;

$\mathrm{\gamma <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>G<br><br>\left(t\right)=\frac{f_1<br><br>\left(t\right)-f_2<br><br>\left(t\right)-f_3<br><br>\left(t\right)+f_4<br><br>\left(t\right)}{4<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>Z}$$\mathrm{\gamma <br><br>\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>B<br><br>\left(t\right)=\frac{f_1<br><br>\left(t\right)+f_2<br><br>\left(t\right)-f_3<br><br>\left(t\right)-f_4<br><br>\left(t\right)}{4}$

where f_n(t) represents time-varying precessing frequency of said spin-echo signal;

Δ

G(t) and Δ

B(t) are, respectively, the EC gradient and B₀ oscillation produced by an eddy current;

γ

is a gyromagnetic ratio value of an appropriate nuclear species; and

quantitative values for precessing frequencies of spin-echo signals are determined from acquired NMR signal data samples according to the following relationships;

$f_n<br><br>\left(i\right)=\frac{u<br><br>\left(i\right)<br><br>v<br><br>\left(i\right)-v<br><br>\left(i\right)<br><br>u<br><br>\left(i\right)}{u^2<br><br>\left(i\right)+v^2<br><br>\left(i\right)}$$\mathrm{with}$$\mathrm{\mu <br><br>}<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)+S_n^u<br><br>\left(i\right)}{2}$$v<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)+S_n^v<br><br>\left(i\right)}{2}<br><br><br><br>{\hspace{0.17em}<br><br>}_{\stackrel{.}{u}}<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)-S_n^u<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}<br><br><br><br>{\hspace{0.17em}<br><br>}_{\stackrel{.}{v}}<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)-S_n^v<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$

where f_n(i) for n=1, 2, 3, 4 represents precessing frequencies of four NMR signals corresponding to said first and second pair of spin-echo signals;

i for i=0,1, 2, . . . N−

2 represents a particular data sample with N being a predetermined number of data samples acquired from a spin-echo NMR signal;

S_n^u and S_n^v are real and imaginary parts of complex NMR signal data acquired;

{dot over (u)}(i) and {dot over (v)}(i) are corresponding time derivatives of u(i) and v(i), respectively; and

Δ

t is a sampling rate for acquiring NMR signal data;

b) pre-distorting a gradient pulse waveform of a predetermined MRI sequence based on a quantified value for the EC gradient determined in step (a) such that a resultant gradient field produced during the sequence is appropriately compensated against the effects of any gradient switching induced eddy current; and

c) producing a counter-oscillating B₀ field based on a quantified value for B₀ oscillation determined in step (a) to yield a stable B₀ field.

14. In an MRI system having a background static magnetic field, B₀, and gradient coils to induce switched gradients in the static field, a method for quantitating background static field gradient, comprising the steps of:

a) placing an MR active test object at a center of an imaging volume of the MRI system and aligning the object along the measurement direction;

b) applying a first gradient pulse to a gradient coil and acquiring signal data samples from a first pair of spin-echo NMR signals corresponding to two slices located symmetrically apart along said measurement direction away from a center of the magnetic field B₀;

c) applying a second gradient pulse to the gradient coil, the second gradient pulse having a polarity opposite to the first gradient pulse, and acquiring signal data samples from a second pair of spin-echo NMR signals corresponding to the slices of step (b);

d) determining four precessing frequencies, f₁-f₄, from acquired signal data samples of each spin-echo signal of said first and second pair of spin-echo NMR signals; and

e) quantifying a background gradient, {overscore (G)}_Z, from the precessing frequencies determined in step (d) according to the following relationship;

${\stackrel{\_}{G}}_Z=\frac{f_1-f_2+f_3-f_4}{4<br><br>\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>Z}.$

15. A method as in claim 14 wherein quantitative values for precessing frequencies of spin-echo signals are determined from acquired NMR signal data samples according to the following relationships:

$f_n<br><br>\left(i\right)=\frac{u<br><br>\left(i\right)<br><br>v<br><br>\left(i\right)-v<br><br>\left(i\right)<br><br>u<br><br>\left(i\right)}{u^2<br><br>\left(i\right)+v^2<br><br>\left(i\right)}$$\mathrm{with}$$u<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)+S_n^u<br><br>\left(i\right)}{2}$$v<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)+S_n^v<br><br>\left(i\right)}{2}$${\hspace{0.17em}<br><br>}_{\stackrel{.}{u}}<br><br>\left(i\right)=\frac{S_n^u<br><br>\left(i+1\right)-S_n^u<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$${\hspace{0.17em}<br><br>}_{\stackrel{.}{v}}<br><br>\left(i\right)=\frac{S_n^v<br><br>\left(i+1\right)-S_n^v<br><br>\left(i\right)}{\mathrm{\Delta <br><br>}<br><br>\text{\hspace{1em}<br><br>}<br><br>t}$

where f_n(i) for n=1, 2, 3, 4 represents precessing frequencies of four NMR signals corresponding to said first and second pair of spin-echo signals;

i for i=0,1, 2, . . . N−

2 represents a particular data sample with N being a predetermined number of data samples acquired from a spin-echo NMR signal;

s_n^u and s_n^v are real and imaginary parts of complex NMR signal data acquired;

{dot over (u)}(i) and {dot over (v)}(i) are corresponding time derivatives of u(i) and v(i), respectively; and

Δ

t is a sampling rate of acquired NMR signal data.

16. A method for evaluating and compensating for large background static field gradients occurring during EC measurements or MRI sequences, comprising the steps of:

a) determining quantitative values of a background field gradient present within an EC measurement area due to B₀ field inhomogeneities and other static field gradients within an MRI apparatus; and

b) applying small opposing field gradients of equal magnitudes along directions corresponding to the background field gradient and other static field gradients.

17. A method as in claim 16 wherein an asymmetric (ASE) MRI pulse sequence is used for obtaining NMR signals and quantitative values for background static field gradient are determined according to the following two relationships: