Electricity: measuring and testing – Particle precession resonance – Using a nuclear resonance spectrometer system
Reexamination Certificate
2000-12-11
2003-05-06
Lefkowitz, Edward (Department: 2862)
Electricity: measuring and testing
Particle precession resonance
Using a nuclear resonance spectrometer system
C324S309000
Reexamination Certificate
active
06559641
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention is directed to a method for operating a magnetic resonance tomography apparatus with a gradient system and a radio-frequency system that, among other things, serve for location coding.
2. Description of the Prior Art
Magnetic resonance tomography is a known technology for acquiring images of the inside of the body of an examination subject. In a magnetic resonance tomography apparatus, rapidly switched gradient fields that are generated by a gradient system are thereby superimposed on a static basic magnetic field. The magnetic resonance tomography apparatus also has a radio-frequency system that emits radio-frequency signals into the examination subject for triggering magnetic resonance signals and picks up the generated magnetic resonance signals. Image datasets are produced on the basis of these magnetic resonance signals. The gradient system and possibly the radio-frequency system are set such that, among other things, they effect a location coding within the examination subject.
All methods that use a repeated scanning of a structure of organs and tissues in order to image processes that change over time, for instance physiological functions or pathological events, are generically referred to as functional imaging in general medicine. In the narrower sense in magnetic resonance tomography, functional imaging means measuring methods that make it possible to identify and image brain areas of a patient that participate in a specific motor, sensory or cognitive task. To this end, three-dimensional datasets of the brain are registered, for example, every two through four seconds, for example with an echo planar method. Echo planar methods have the advantage that the image dataset registration required for a single three-dimensional dataset is obtained very fast, in less than 100 ms.
An image dataset thereby contains a number of picture elements that, with a suitable arrangement, generate an image in a grid-like raster. A picture element, for example, is characterized by a value on a gray scale as well as by its coordinates, i.e. its position within the image. For a two-dimensional image, for example, the picture elements are entered in a two-dimensional matrix that forms a two-dimensional image dataset. The arrangement within the matrix thereby defines the coordinates of the picture elements in the image. The analogous case applies to a three-dimensional image dataset.
After many image datasets with coinciding location coding have been registered at different points in time within the framework of a functional magnetic resonance tomography procedure, image datasets are subtracted from one another for forming images referred to as activation images, i.e. the datasets compared to one another in terms of signal differences for the identification of active brain areas. Even the slightest movements of the brain during the entire registration time span produce undesired signal differences that mask the brain activation being sought.
For filtering out the aforementioned undesired signal differences, correction referred to as a retro-perspective motion correction of the image datasets is first implemented, which assumes an acquisition of positional changes as a result of movements. A method for this is based, for example, on the assumptions that movements occur only between the registrations of individual, complete image datasets, and that the brain can be considered to be a rigid body. Further, an arbitrary rigid body movement in three-dimensional space is described with six motion parameters, whereby three parameters identify translations and three parameters identify rotations.
The aforementioned parameters are, for example, noted in a column vector {right arrow over (q)}. The values of all picture elements or of selected picture elements of a first image dataset and of a second image dataset that was generated temporally following the first are noted in a coinciding sequence in a first column vector {right arrow over (x)} and a second column vector {right arrow over (y)}. For determining a positional change between the registration times of the first and of the second image dataset, i.e. for determining the motion parameters, the following equation (which is based on a Taylor expansion of the 1
st
order) is solved, for example by an iterative method:
y
_
-
x
_
=
[
∂
x
1
∂
q
1
⋯
∂
x
1
∂
q
6
⋮
⋰
⋮
∂
x
n
∂
q
1
⋯
∂
x
n
∂
q
6
]
·
q
-
with
x
_
=
[
x
1
⋮
x
n
]
;
y
_
=
[
y
1
⋮
y
n
]
;
q
_
=
[
q
1
⋮
q
6
]
The above equation contains a Jacobean functional matrix that has partial derivatives of the elements of the column vector {right arrow over (x)} according to the six motion parameters per row. The book by R. S. J. Frackowiak et al.,
Human Brain Function,
Academic press, 1997, particularly Chapter 3, pages 43-58, is referenced for a more detailed description of the aforementioned method for determining positional changes from image datasets. The article by U. Klose et al., “Funktionelle Bildgebung mit der Magnetresonanztomographie”, elektromedica 67 (1999) Number 1, pages 27-36, can be consulted for a more detailed description of the procedure in functional magnetic resonance tomography with a retro-perspective motion correction of image datasets that occurs temporally following a complete registration of all image datasets of functional magnetic resonance tomography.
Several methods for retro-perspective motion correction assume that all undesired signal differences as a result of movements can be eliminated given an optimum retro-rotation or retro-shift of the image datasets with respect to a reference image dataset. Often left out of consideration, however, is that all undesired signal differences caused by movements cannot be filtered out with a mere geometrical shifting, or rotation of the image datasets. The reason for this is that, following a change in position of the brain, gradient fields and radio-frequency fields act differently on specific volume regions of the brain compared to the initial position given unaltered location coding, and thus excitation, resonance and relaxation properties of these volume regions change. As a result, the signal behavior of these volume regions is modified not only for an immediately successively registered image dataset but also persistently for further image datasets to be registered. If a volume region under consideration changes in position repeatedly during the overall functional magnetic resonance tomography, then each change in position has an influence—as a signal modification—more or less up to the last image dataset that is registered.
With respect to this latter signal modification and, for example, given a slice-by-slice registration of three-dimensional image datasets, those positional changes are particularly important for slices that are rotated or shifted in planes other than their own plane. This, for example, is described in greater detail in the article by K. J. Friston et al., “Movement-Related Effects in fMRI Time-Series”, Magnetic Resonance in Medicine 35 (1996), pages 346-355. This article, moreover, proposes an approximation method with which the latter, motion-caused signal differences also can be filtered out of the image datasets following the production of image datasets in functional magnetic resonance tomography. Only very limited corrections are possible, however, with this approximation method.
Another approach for avoiding all undesired, motion-caused signal differences is, instead of correcting all image datasets in the context of a prospective motion correction afterwards, to acquire potential positional changes of the brain from image dataset to image dataset during the execution of a functional magnetic resonance tomography procedure and to appropriately adapt the location coding during the run.
In one embodiment of this known approach, positional changes of the head are optically
Lefkowitz Edward
Schiff & Hardin & Waite
Siemens Aktiengesellschaft
Vargas Dixomara
LandOfFree
Method for operating a magnetic resonance tomography... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for operating a magnetic resonance tomography..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for operating a magnetic resonance tomography... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3073618