Picker International, Inc.

Emi, LLC

Michael Burl

Ian R Young

Colin G Harrison

Godfrey N Hounsfield

Hugh Clow

Graham J Clarke

A method and apparatus for determining the rate of flow of a liquid in a selected region of a body by nuclear magnetic resonance techniques. The method includes a sequence of applying a first magnetic pulse (B.sub.1 (90.degree.)) effective to excite nuclear magnetic resonance of a chosen nucleus within the liquid preferentially in a slice of the body which includes the selected region. A period of time (t.sub.D) is waited and then a second magnetic pulse (B.sub.1 (90.degree.)) is applied which is effective to excite nuclear magnetic resonance of the nuclei preferentially in the slice, and the free induction decay signal is measured. The whole sequence is repeated for different values of the period of time (t.sub.D). The values of the period of time (t.sub.D) and the corresponding measured signals are then related to the rate of flow of the liquid through the slice.

Nuclear magnetic resonance method and apparatus

Nuclear magnetic resonance methods have been proposed for examination of large bodies, including medical examination of patients. These include so-called steady state free precession methods. Such methods involve alternating excitation of a region of the body in such a way that an equilibrium condition is established at which a mean resonance signal is detectable. Typically for excitation of a planar region a radio frequency excitation field is provided in the presence of a field having an alternating gradient so that equilibrium is only reached in the plane. It is now proposed to achieve the same result by the application of two radio frequency fields having opposing gradients and pulsed in alternation to balance substantially only in the plane and provide equilibrium only there. The fields are provided by respective sets of saddle shaped coils divided into pairs of unequal dimension (1a, 1b, 2a, 2b) to provide the respective gradients.

Nuclear magnetic resonance systems

In nuclear magnetic resonance (NMR) imaging systems it is known to excite resonance in a slice of a body and then to sample the resonance signals in the presence of a field gradient across the slice. It has been proposed that the signals should be sampled at intervals such that there is equal field integral (for the field gradient) in each interval. It is now proposed to use outputs from two NMR probes in the slice and on opposite sides of the body. The phase difference between the two probes is measured and samples are taken when the phase difference reaches a predetermined value. Preferably the samples are taken when a gate detects zero-crossing of the output of a demodulator providing the phase difference. Two probes only may be rotated around the patient as the field gradient rotates or four may be used, two orthogonally disposed pairs and one or other appropriately weighted output used.

Nuclear magnetic resonance apparatus

The invention provides RF field coils and detector coils for an NMR machine. The two coils are specially designed for their respective purpose, the transmit coil (5, 6) being rectangular and the received coil (7-11) substantially elliptical with defined axis ratios and spacing in the z-direction being by projection of equi-spaced conductors on to the ellipse.

Imaging systems

The invention provides an investigation of chemical shift for an element, for example phosphorus, in a region of a body, in one example resonance is preferentially in a line in a slice of the body. Frequency dispersion down the line is produced by a pulsed field having a switched gradient so that the frequency distribution is in steps. This allows a chemical dispersion in each limited region of the line without overlap with adjacent regions. The measured FID signals are Fourier transformed to give a spatial and chemical analysis of the line. Alternatively resonance is excited in a slice and lines selected by a pulsed gradient which gives chemical dispersion as well as spatial dispersion between the lines after Fourier Transformation. This is repeated for many different directions of the pulsed gradient to allow analysis of a chemical line by convolution techniques to give a picture for the slice.

The invention is related to the so-called 'echo-planar' method of NMR imaging. In that method a resonance is first excited in a slice and then two orthogonal gradients provided dispersion in the slice. One gradient is pulsed to space the frequencies of spins in adjacent strips allowing the other gradient to provide dispersion down the strips. It is now proposed to pulse the second gradient so that the dephasing down each strip is also in steps. The FID signals (including rephasing signals) for each distinct element so produced can then be put in an array and two-dimensionally Fourier transformed to yield the desired output for each element.

In known NMR imaging apparatuses it has been the practice to arrange the G.sub.R field gradient, which rotates in the plane of the examined slice, with a zero crossing substantially in the center of the slice and to demodulate at the Larmor frequency at the zero crossing. It is now proposed to demodulate the detected resonance signals at a frequency which, in the presence of the field gradient, is for a part of the plane lying outside the body slice. This may be by arranging the zero-crossing to be outside the slice. In that case it is preferable that the zero-crossing of G.sub.R lies on a straight line (27) maintained tangential to a circle (28) lying outside the body (23) about an axis (24) through the slice. Control may then be by precalculated coil currents or by precalculating fields at probe positions and adjusting the coil currents to give the expected fields. Alternatively the zero-crossing may be maintained at the center and demodulation may be at a different frequency. It is then desirable also to measure the resonance frequency in the absence of the field gradient across the slice. The resonance signals are then effectively further demodulated at that measured frequency. The effective demodulation is preferably achieved by complex multiplication of the two signals.

In a medical NMR apparatus it is desirable that the axial magnetic field, Ho, is uniform. The result can in theory be obtained by accuracy of construction but in practice this cannot readily be achieved. It is now proposed to use a field sensing probe system which indicates the value of that field at a plurality of positions in the examined region and provides error signals indicating deviations from uniformity. The field is automatically corrected in response to the error signals by applying currents to field correcting coils. The errors may be considered as distinct gradient and dishing errors which are corrected individually.

Growing community of inventors

Iver, United Kingdom

Michael Burl