WO2014024114A1

WO2014024114A1 - A data transfer circuit, method and system for an mri machine having a plurality of receiver surface coils

Info

Publication number: WO2014024114A1
Application number: PCT/IB2013/056403
Authority: WO
Inventors: Marius Eugene GOOSEN; Monuko Du Plessis; Jan Gerrit KORVINK
Original assignee: Insiava (Pty) Ltd.; Albert-Ludwigs-Universität Freiburg
Priority date: 2012-08-06
Filing date: 2013-08-05
Publication date: 2014-02-13
Also published as: ZA201501454B

Abstract

The invention is for a data transfer circuit (10), method (100) and system (140) with the ability to receive and process plural input signals by a single data transfer circuit with on-die signal combining and optical modulation. The circuit (10) is for an MRI machine having a plurality of receiver surface coils (12), and includes a plurality of input connections (14), each input connection (14) electrically connected or connectable to a receiver surface coil (12) thereby to receive an electrical analogue input signal from the receiver surface coil (12). The circuit (10) has at least one combiner (26, 36) operable to combine the plural input signals received from the input connections (14) into an electrical analogue combined signal (26.1, 36.1 ) and an optical driver (38) and light emitter (40) operable to modulate the combined signal (26.1, 36.1) into an optical analogue output signal (40.1), thereby to transmit the output signal (40.1) optically to a remote data processing unit (160).

Description

A DATA TRANSFER CIRUIT, METHOD AND SYSTEM FOR AN MRI MACHINE HAVING A PLURALITY OF RECEIVER SURFACE COILS

FIELD OF INVENTION

The present invention relates in general to magnetic resonance imaging (MRI) systems employing a plurality of receiver surface coils, and more particularly to a data transfer circuit, method and system for an MRI machine having a plurality of receiver surface coils for efficient data transfer of a plurality of analogue MR data signals received from the plurality of receiver surface coils to a data processing unit.

BACKGROUND OF INVENTION

Nuclear MRI relies on the quantum mechanical property of atomic spin. Nuclei with non-zero uneven spin have a magnetic moment which when subjected to a magnetic field can polarize into one of two or more energy states, the energy difference between the states being related to the Planck constant. On a macroscopic level, the alignment of the protons creates a net magnetisation vector (longitudinal magnetisation) in the direction of the main magnetic field (B₀-field). The strength of the magnetisation is directly proportional to the strength of the B₀-field according to a Boltzmann distribution among the available energy states.

With the application of a radio frequency (RF) pulse in a direction that is perpendicular to the Bo-field, a nucleus can resonantly absorb and emit the transmitted energy enabling it to move among the energy states (e.g., against the direction of the B₀-field), and macroscopically, the magnetisation will rotate away from the direction of the Bo- field. The moment the magnetisation rotates away from the direction of the B₀-field, it starts precessing at the Larmor frequency of the nucleus. The exact frequency of precession depends sensitively on the molecule type situated in the measurement volume. With the removal of the radio frequency pulse, the magnetisation will continue to precess, and the protons will start to lose magnetisation and/or coherency through either spin-spin interactions or spin-lattice interactions. This reduction results in a realignment of the magnetisation along the B₀-field with time. The strength of the magnetisation, and its rates of decay hence translates to a quantitative property which can be used to create for example an MR-image. [E. L. Hahn and P. Mansfield, "NMR and MRI in Retrospect [and Discussion]," Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 333, no. 1632, pp. 403-411 , Dec. 1990.] With the application of additional gradient fields (G_x, G_y, G_z), position is encoded in the decaying signals as a function of frequency and phase. The choice of radio frequency pulse and gradient field timing depends on the scanning sequence being implemented. The position encoded decaying signals are predominantly and typically Faraday- detected by means of an RF-coil tuned to the Larmor frequency of the nucleus being used for imaging under the specific B₀-field strength, although other detection mechanisms exist.

Coil arrays provide extra flexibility with improved signal to noise ratio (SNR) while reducing the scan time. This is due to the fact that smaller receiver coils are more sensitive to an area directly adjacent to it, and have less resistive noise, thereby improving the SNR in the smaller volume. The receiver coils, typically surface spiral coils but also surface solenoidal coils [O. G. Gruschke, N. Baxan, L. Clad, K. Kratt, D. von Elverfeldt, A. Peter, J. Hennig, V. Badilita, U. Wallrabe, and J. G. Korvink, "Lab On a Chip Phased-array MR Multi-platform Analysis System," Lab Chip, vol. 12, no. 3, pp. 495-502, Dec. 2011.], are typically arranged in an overlapping function to reduce the mutual inductance to nearest neighbouring coils. Since each overlapping coil has a unique location relative to the volume being imaged, each coil contains spatial information which allows the amount of magnetic field gradient steps to be reduced, in particular the phase encoding steps. This reduces the scan time while achieving the SNR of a small surface coil. Hence combining a plurality of these surface coils, spread about the total subject to be imaged, increases the overall SNR and reduces scan time at a cost of increased interface complexity and data processing, and by operating these coils as a phased array, the overall SNR can be further improved [P. Roemer, W. Edelstein, C. Hayes, S. Souza, and O. Mueller, "The NMR Phased-array," Magnetic Resonance in Medicine, vol. 16, no. 2, pp. 192-225, 1990.].

Most widely used MR or MRI systems utilise electrical cables, such as shielded coaxial cables, to transfer the MR-signals from the sensing coils to the data processing unit. The cables however are subjected to high switching fields (G_x, G_y, G_z) inducing sheath waves and degrade the system SNR. These cables also do not provide a scalable solution for massively parallel surface coil arrays due to the bulkiness of the cables themselves as well as the required resonant traps to reduce the effect of the noisy sheath waves. Not only the cables, but also the connectors used to make electrical contact, introduce unwanted noise in the system. Therefore, the introduction of a plurality of surface coils increases the achievable SNR, or alternatively maintains the same SNR at a reduced scan time, but the method of relaying the signals to the data processing unit deteriorates it again, practicable solutions thus usually having a small number of coils. In theory, an optimal number of surface coils approaches a very large number.

A number of alternatives to copper cables have been proposed, for example using modulators to modulate an incoming laser according to the digitised MR-signal to a wireless radio link. However each of the alternative solutions lack in cost effective scaling of a massively parallel surface coil array.

The resolution of an MR-image can theoretically be improved with an increase in the amount of coils, however, reducing the coil diameter and maintaining a high quality factor (Q-factor) receiver coil remains challenging. It would therefore be desirable to have a compact integrated system which could combine analogue MR-signals from a plurality of receiver coils and transfer the data optically to a data processing unit. This would allow a reduction in cabling needs, increasing scalability of multiple coil receiver systems, while maintaining a high SNR ultimately to approach a one-voxel-one-coil system.

SUMMARY OF INVENTION

According to a first aspect of the invention, there is provided a data transfer circuit for an MRI machine having a plurality of receiver surface coils, the data transfer circuit including:

a plurality of input connections, each input connection electrically connected or connectable to a receiver surface coil thereby to receive an electrical analogue input signal from the receiver surface coil;

at least one combiner operable to combine the plural input signals received from the input connections into an electrical analogue combined signal; and an optical driver and light emitter operable to modulate the combined signal into an optical analogue output signal, thereby to transmit the output signal optically to a remote data processing unit. The data transfer circuit may be an integrated single-die circuit and the light emitter may be an on-die light emitter.

The data transfer circuit may be a silicon circuit and the light emitter may be a silicon light emitter.

The data transfer circuit may include a converter associated with each input connection operable to convert the received input signal to an intermediate signal having an intermediate frequency. The converter may be a downconverter or an upconverter thereby operable to downconvert or upconvert the received input signal.

In one embodiment, the data transfer circuit may include plural stages of converting and multiplexing. The plural input signals (referred to as a group of signals) may be divided into sub-groups, the data transfer circuit including a plurality of first stage converters (e.g. downconverters) operable to convert (e.g. downconvert) each signal within a sub- group to a unique intermediate frequency, but corresponding signals in different subgroups having the same intermediate frequency.

Thus, the data transfer circuit may include:

a plurality of first stage multiplexers, with a single multiplexer being associated with each sub-group of signals, each multiplexer being operable to combine the converted signals of each sub-group into a sub-group signal ;

a second stage converter operable to convert each sub-group signal to a unique intermediate frequency; and

a second stage multiplexer operable to combine the converted sub-group signals into a combined group signal.

The group signal may be plural and transferred using coarse wavelength division multiplexing, thereby increasing scalability. The individual converted input signals in the combined group signal may be spread over a chosen bandwidth with a chosen channel spacing, such that neighbouring individual signals will not interfere with each other for a pre-defined amount of frequency shift. The data transfer circuit may include a filter associated with each converter thereby to remove any additional frequency components and leave only the converted frequency component.

Using the fact that the MR-signal data, bearing in mind the decay of the transversal and longitudinal magnetisation vectors, results in a low bandwidth data signal situated at a higher carrier, a multitude of these signals can be combined in the frequency domain without affecting each other. Chemical shifts, non-ideal coil-to-subject coupling and magnetic field inhomogeneity result in a carrier shift, hence provision may be made for such shifts by spacing the data bands in the spectrum. By keeping the signal in the analogue domain in the data transfer circuit, unwanted noise, introduced by switching of analogue to digital conversion circuits, is removed, indicating a possible SNR increase, ultimately resulting in an improved image quality.

The data transfer circuit may include a tuning, matching and detuning circuit, in no specific order, and a low noise amplifier associated with each input connection. The tuning circuit adapts the resonant frequency of the surface coil to match the Larmor frequency of the nucleus to be imaged, while a matching circuit changes the perceived impedance of the coil to match the low noise amplifier for optimal signal transfer. The detuning circuit, in turn, changes the tuned resonance frequency of the coil upon transmission of the excitation pulse in order to protect the low noise amplifier. The light emitter may be integrated in the same low cost complementary metal oxide semiconductor (CMOS) process used to implement the low noise amplifier of each of the associated receiver coils, improving the scalability by removing the need for external components. The scalability of a massively parallel high resolution MR-system is an important factor, and hence an amount of MR-signals may be combined to reduce the amount of fibre optic connectors as well as the amount of fibre optic cables.

The data transfer circuit may include a resonator tuned to an RF excitation pulse frequency of the MRI machine thereby to provide on-chip power generation. The data transfer circuit may include a non-ferrous electrically conducting enclosure or cage, operable to shield electrical fields. The enclosure may define an aperture aligned with the output connection thereby to align an optical fibre and the optical output connection in use.

According to another aspect of the invention, there is provided a data transfer circuit for an MRI machine having a plurality of receiver surface coils, the data transfer circuit including:

a plurality of input connections, each input connection electrically connected or connectable to a receiver surface coil thereby to receive an electrical input signal from the receiver surface coil;

at least one combiner operable to combine the plural input signals received from the input connections into an electrical combined signal; and

an on-die optical driver and light emitter operable to modulate the combined signal into an optical output signal, thereby to transmit the output signal optically to a remote data processing unit,

the circuit being an integrated single-die circuit.

The integrated circuit may be a silicon circuit and the light emitter may be a silicon light emitter.

The invention extends to a data transfer method for an MRI machine having a plurality of receiver surface coils, the data transfer method including:

receiving, by an integrated circuit, a plurality of electrical analogue input signals from the receiver surface coils;

combining the plural input signals into an electrical analogue combined signal; and

modulating the combined signal into an optical analogue output signal by means of a light emitter.

The method may be implemented on an integrated single-die silicon circuit and in which the light emitter is an on-die silicon light emitter. The method may include converting each received input signal to an intermediate signal having an intermediate frequency prior to combining. The method may include plural stages of converting and combining. The invention extends further to a data transfer system for an MRI machine having a plurality of receiver surface coils, the data transfer system including:

at least one data transfer circuit as defined above;

a plurality of receiver surface coils operatively connected to the input connections of the data transfer circuit; and

an optical fibre operatively coupled to the light emitter of the data transfer circuit.

The data transfer system may include a plurality of data transfer circuits for large scalability. The optical fibre from each circuit may be grouped into a fibre bundle for optical transmission to a remote data processor including an optical receiver, separator and Analogue to Digital Converter (ADC). While the data processor may be remote, it is typically not far from the data transfer circuit, thus maintaining good SNR characteristics. Utilising a fibre optic cable as transmission medium further allows the data processor to be situated further away without degrading the system SNR.

The data transfer circuit may be a bulk CMOS process. However, the data transfer circuit, and particularly the light source array used for optical transmission of the MR- signals, may not necessarily be limited to bulk CMOS processes but may include other processes allowing the use of an integrated light source. Other processes include, but are not limited to, SOI-CMOS and Si:Ge.

The semiconductor composition of the data transfer circuit may be based on, include aspects of, or be adapted from the devices described in US patent nos. 5,994,720 and 6,1 1 1 ,271 respectively (but it is not limited thereto). US5994720 and US61 1 1271 are hereby incorporated by reference. BRIEF DESCRIPTION OF DRAWINGS

The invention will now be further described, by way of non-limiting example, with reference to the accompanying diagrammatic drawings.

In the drawings:

Figure 1 shows a schematic circuit diagram of a data transfer circuit in accordance with the invention;

Figure 2 shows a schematic cross-sectional view of one embodiment of the circuit of Figure 1 ;

Figure 3 shows a schematic circuit diagram of a power-generating portion of another embodiment of the circuit of Figure 1 ;

Figure 4 shows a diagrammatic representation of signal outputs of the circuit of

Figure 1 ;

Figure 5 shows a flow diagram of a data transfer method in accordance with the invention;

Figure 6 shows schematic views of example coil layouts for use with the circuit of

Figure 1 ;

Figure 7 shows a schematic view of a data transfer system in accordance with the invention; and

Figure 8 shows a schematic view of the data transfer system of Figure 7, which has been up-scaled for an increased number of surface coils.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENT

Referring initially to Figure 1 , reference numeral 10 generally indicates a data transfer circuit in accordance with the invention. The data transfer circuit 10 is for use in an MRI machine or MR-machine (not separately illustrated) having a plurality of receiver surface coils 12 which are receptive to a generated MR-signal after being perturbed by an RF excitation pulse. (See Figure 6 for a more thorough description of the receiver coils 12.) In broad terms, the data transfer circuit 10 includes input connections 14, at least one combiner 26, 36 and an optical driver 38 and a light emitter 40, but these elements are described in more detail below.

More specifically, the data transfer circuit 10 has a plurality of input connections 14, each input connection 14 electrically connected to a receiver surface coil 12. The data transfer circuit 10 thus has a one-coil one-input configuration. The input connections 14 are in the form of electrical contacts (see Figure 2) and are able to receive respective electrical analogue input signals from the receiver surface coils 12. A tuning circuit 18.1 , which changes the resonant frequency of the coil to match the Larmor frequency of the nucleas to be imaged, a matching circuit 18.2, which alters the impedance of the coil for optimal coupling to the next stage, a detuning circuit 18.3, which changes the coil resonant frequency when an RF excitation pulse is transmitted, and a low noise amplifier (LNA) 20 are electrically associated with each input connection 14. The tuning circuit 18.1 , matching circuit 18.2 and detuning circuit 18.3, in no specific order, constitute the interface circuitry 18 between the surface coil 12 and the LNA 20. An electrical input 1 6 directs the detuning circuit 18.3 when to be active or inactive, changing the resonant frequency of the tuned coil in order to protect the LNA 20.

The data transfer circuit 10 has a plurality of converters 22, 32 and a plurality of combiners 26, 36. This embodiment illustrates a plural stage data transfer circuit 10 having two stages of signal converting and combining (refer also to Figure 5 and the associated description). In broad terms, each converter 22, 32 is operable to convert the received input signal to an intermediate signal having an intermediate frequency. Similarly, each combiner 26, 36 is operable to combine plural input signals (i.e. the converted intermediate signals) into an electrical analogue combined signal. In this embodiment, the converters 22, 32 are downconverters and the combiners 26, 36 are multiplexers. Downstream of each downconverter 22, 32 is a filter 24, 34 to remove any additional frequency components (other than the desired intermediate frequency). The final multiplexer 36 feeds the optical driver 38 which, in turn, drives the light emitter 40. The optical driver 38 and light emitter 40 are operable to modulate the combined signal into an optical analogue output signal, thereby to transmit the output signal optically to a remote data processing unit (refer to Figures 7 and 8). In this example embodiment, the data transfer circuit 10 is an integrated single-die silicon circuit and the light emitter 40 is an on-die silicon light emitter. This single-die configuration is advantageous in that it provides excellent SNR characteristics for analogue data transfer as described. However, if practicable, the data transfer circuit could comprise two or more discrete components (not illustrated). A single-die configuration is illustrated in Figure 2. The data transfer circuit 10 could have on-chip power generation (refer to Figure 3) or a power supply (not illustrated). It will be appreciated by one skilled in the art that some form of power supply is required (whether on-chip or external) and that the power supply is not illustrated in every Figure.

In one embodiment (as illustrated in Figure 2), the data transfer circuit 10 includes a Faraday cage or enclosure 50 which provides electrical shielding to the circuit die 54 therein. In an MRI environment ferrous metals may not be used due to the extreme magnetic field and the disturbance to the MR detection process (B₀-field). Any ferrous metals will be ripped from their fixation and accelerated in the vector direction of the Bo- field, damaging the MRI machine. Even if strapped down, ferrous metals will disturb the exquisite homogeneity of the B₀-field, which in turn will destroy the MR detection process, ultimately distorting the image to be formed. Since only non-ferrous metals can be used, no complete magnetic field shielding can be accomplished and magnetic field effects such as the Hall-effect should be taken into account in the RF front-end design. Thus, the cage 50 is of aluminium which provides some useful electrical shielding but is not itself strongly susceptible to the surrounding magnetic fields. The cage 50 is fast with a printed circuit board (PCB) 56 on which the die 54 is mounted. A bottom layer of the PCB 56 is utilised as a ground plane 64 as far as possible to accomplish shielding from the bottom plane. The ground 64 is of a non- ferrous electrically conducting material which includes the likes of aluminium, copper and gold. This allows a complete surrounding of the die 54 with a non-ferrous conducting metal.

The cage 50 defines therein a fibre aperture 52 for locating an optical fibre and aligning it with the light emitter (i.e. an optical output) 40 of the die 54. Lenses or other optical techniques can be utilised to increase the fibre coupling efficiency. A diameter of the aperture 52, through which a fibre ferrule fits, should typically be less than a 10^th of a quarter wavelength of the frequency being shielded. Hence, in a worst case 9.4 T system, with a Larmor frequency of 400 MHz, a diameter of 1 .875 mm can be safely made without affecting the shielding capability. The IC die 54 is usually wire bonded in the centre of a PCB, repeatable to within a few micrometers. Manufacturing the aluminium cage 50 to be self aligning with the PCB 56 allows the fibre to be self-aligned with the source 40 situated in the centre of the IC die 54.

The circuit 10 also includes aluminium or gold bond wires 60 connected to the input connections 14 which may be of aluminium, copper or gold.

In another embodiment, as illustrated in Figure 3, the circuit 10 includes on-chip power generation. An inductor, for example a spiral, wire or flat solenoidal inductor, 80, tuned to the RF excitation pulse frequency of the MRI machine, is implemented for example on a top metal layer of the circuit 10 for reception of energy in the RF excitation pulse. The alternating RF pulse induces an alternating signal in the tuned inductor 80 and hence rectification is necessary to reach a stable DC level. A cascaded rectifier 82 utilises a few stages to boost the received voltage, acting as a charge pump. The DC energy harvested can be temporarily stored in a capacitor 84 for use after the excitation pulses are transmitted. The DC voltage level stored on the capacitor 84 should be regulated by a regulator 86 in order to be used as the standard process voltage level. Using a 0.35 μηι process, the incoming DC value should be regulated to 3.3 V for standard CMOS level circuits without risking transistor breakdown.

Reference is now made to Figures 1 , 4 and 5 to describe the data transfer circuit 10 in use. Figure 4 shows schematic signals and Figure 5 shows a flow diagram 100 of a method implemented by the circuit 10. A plurality of electrical analogue input signals are received (at block 102) by the electrical inputs 14 from the receiver surface coils 12. The input signals are detuned and amplified (at block 104). The circuit 10 of Figure 1 shows a two stage filter/combine configuration. If desired, this could be scaled down to a single stage only or scaled up to three or more stages.

Each circuit 10 has a group of inputs 14 which are divided into sub-groups. The first stage downconverter 22 is arranged downstream of the LNA 20 and is operable to downconvert (at block 106) each input signal within a sub-group to a unique intermediate frequency 22.1 . These first stage intermediate frequencies 22.1 may be represented by f-i , -- f_n- However, corresponding signals 22.1 in different sub-groups have the same intermediate frequency. For example, a first intermediate signal 22.1 within each sub-group will have the frequency f-i , while the second intermediate signal 22.1 within each subgroup will have the frequency f₂, and so forth.

Each first stage multiplexer 26 is associated with a particular sub-group and is operable to combine (at block 108) each intermediate signal 22.1 with a sub-group into a combined sub-group signal 26.1 . The multiplexers 26 may combine the signals 22.1 in the frequency domain by simple addition.

Similar steps are followed in the second stage. The converters 32 (which could be upconverters or downconverters but in this embodiment are downconverters) downconvert (at block 1 10) each sub-group signal to a unique intermediate frequency f_A, fe... fz to produce a second stage intermediate signal 32.1 . The second stage multiplexer 36 combines (at block 1 12) the converted sub-group signals into a single combined group signal 36.1 . Hence, and as Figure 4 illustrates, in the combined group signal, the subgroups form successive groups of bands, with each converted input signal 22.1 equally spaced in each sub-group, and each sub-group equally spaced within the group. In this fashion, the selected spectrum is filled up in narrow bands which personify the narrowband MR-signals from each receiver surface coil 12. The spectrum and channel spacing is selected such that neighbouring individual signals will not interfere with each other for a pre-defined amount of frequency shift. The complex signal 36.1 is used directly to modulate (at block 1 14) the light source 40 thereby to produce an optical output signal 40.1 for low noise transmission. The output 36.1 of the final multiplexer 36 may still be a plural electrical connection in order to drive the light emitter/s 40 using coarse wavelength division multiplexing, thereby further increasing the scalability.

It will be noted that the total number of receiver surface coils 12 which can be accommodated by a single circuit 10 is the product of the number of converters in the first stage and the number of converters in the second stage. For example, if n = 4 (i.e. four inputs 14 per sub group) and Z = 4 (i.e. four sub-groups), the circuit 10 would accommodate at total of 1 6 receiver surface coils 12.

Referring now to Figure 6, the receiver surface coils 12, which should ideally be situated as close as possible to the area to be imaged, can be configured to establish an array of these coils fitting a non-Gaussian shape through the use of various polygons. Reference is made to US patent no. 7,663,367. As described in US7663367, polygons, as indicated by reference numeral 120, fit together to form a honeycomb structure which can be bent to form a cylinder. Using each of these polygons as a centre-point for a coil 12, with all coils 12 overlapping with nearest neighbours to reduce the mutual inductance, results in an array of coils 122. To implement an additional bend or an increased gradient bend, a polygon with fewer sides can be utilised, forcing the honeycomb structure to assume the required shape. Each of the polygons can be implemented as a separate PCB, with tracks on the PCB forming the surface coil 12 itself. Reference numeral 124 indicates a PCB with the segments of surrounding surface coils 12. Interconnecting each of these PCBs 124, as indicated by reference numeral 126, and jumping the gaps with capacitors, which is used for coil tuning, result in the surface coil array 122. Ideally, for reception of an RF signal, in this case the H relaxation, the coil size should be comparable to the wavelength transmitted. Since this is not an option when implementing a multitude of coils 12, each coil 12 should hence be tuned to resonate at the Larmor frequency, determined by the B₀-field strength and molecule to be imaged, in order to receive the near field signal. Thus, this coil array 122, with a plurality of coils 12, is well suited for use with the data transfer circuit 10.

Referring now to Figures 7 and 8, reference numeral 140 generally indicates a data transfer system in accordance with the invention. The system 140 in Figure 7 includes a single circuit 10, for simplicity of explanation, while the system 140 of Figure 8 includes plural circuits 10 to illustrate the scalability of the circuit 10. Although only two circuits 10 are illustrated, it will be appreciated by one skilled in the art that the system 140 is massively scalable by the addition of more and more parallel circuits 10.

In the system 140, a plurality of MR-signals are received by the sensing surface coils 12, and are combined by the circuit 10 in a scalable fashion and then transmitted across an optical fibre 150 to a spectrometer in an equipment room 1 62 by the integrated silicon light source 40. The scalability of such a system 140 is of extreme importance in enabling the implementation of a large number of receiver coils 12, while keeping the total system 140 manageable and improving the SNR. Since a multitude of coils 12 are connected to each circuit 10, the total number of implemented coils 12 increases rapidly when connecting single fibres 150 into large fibre bundles 151 . The length of the optical fibres 150 is usually short (in relative optical transmission terms) and little or no degradation of the optical signals occurs, thus maintaining a good SNR.

The equipment room 1 62 is configured to receive the optical signals from the plural fibres 150 and therefore includes an optical receiver 152, a LNA 154, and a plurality of separators (in the form of demultiplexers) and ADCs 156 effectively to completely or partially reverse the converting and multiplexing done by the circuit 10, thereby to extract the signals (or least transmitted versions of the signals) from the receiver coils 12. The received signals are then communicated to a data processing unit 1 60 for processing where further signal separation can be performed should the separators 156 only partially demultiplex the compounded optical signal 40.2 received in the equipment room 1 62.

The Applicants believe that the invention as exemplified has a number of advantages and benefits. Importantly, the ability to receive and process plural input signals by a single data transfer circuit 10, and the potential use of plural circuits 10, renders the system 140 massively scalable in terms of the number of input signals 14 from receiver coils 12 which can be handled. The on-die combining and optical modulation of the input signals 14 is well suited to an MRI environment and can be realised with a manageable SNR. The circuit 10 is also compact compared to existing MRI data transfer mechanisms. Furthermore, by increasing the number of plural signals, MRI will benefit from better spatiotemporal resolution, resulting in more accurate MRI images, better time-resolved MR imaged processes, better functional MRI images, and better MRI microscopy images. Furthermore, since MRI is used in many different applications, ranging from medicine to technology, a large user community stands to benefit from this improvement. Silicon is fairly immune, or at least less susceptible than some other materials, to the large magnetic fields in this environment. Depending on the implementation, the circuit 10 can be configured for on-chip power generation, or can include a cage or enclosure 50 to provide electrical shielding and to align the optical fibre 150 with the light emitter 40. The use of an integrated silicon system will increase the ease of manufacturing of such MRI receiver arrays.

Finally, and usefully, the circuit 10 is more cost-effective than currently available conventional MRI data transfer mechanisms.

Claims

CLAIMS:

1. A data transfer circuit for an MRI machine having a plurality of receiver surface coils, the data transfer circuit including:

a plurality of input connections, each input connection electrically connected or connectable to a receiver surface coil thereby to receive an electrical analogue input signal from the receiver surface coil; at least one combiner operable to combine the plural input signals received from the input connections into an electrical analogue combined signal; and

an optical driver and light emitter operable to modulate the combined signal into an optical analogue output signal, thereby to transmit the output signal optically to a remote data processing unit.

2. The data transfer circuit as claimed in claim 1 , in which:

the circuit is an integrated single-die circuit; and

the light emitter is an on-die light emitter.

3. The data transfer circuit as claimed in claim 1 or claim 2, in which:

the circuit is a silicon circuit; and

the light emitter is a silicon light emitter.

4. The data transfer circuit as claimed in any of the preceding claims, which includes a converter associated with each input connection operable to convert the received input signal to an intermediate signal having an intermediate frequency.

5. The data transfer circuit as claimed in claim 4, in which the converter is a down- or upconverter thereby operable to down- or upconvert received input signal.

6. The data transfer circuit as claimed in claim 4 or claim 5, in which the combiner includes a multiplexer operable to combine the plural converted signals into a combined signal. The data transfer circuit as claimed in claim 5 or claim 6, which includes plural stages of converting and multiplexing.

The data transfer circuit as claimed in claim 7, in which the plural input signals (referred to as a group of signals) are divided into sub-groups, the data transfer circuit including a plurality of first stage up- or downconverters operable to up- or downconvert each signal within a sub-group to a unique intermediate frequency, but corresponding signals in different sub-groups having the same intermediate frequency.

The data transfer circuit as claimed in claim 8, which includes:

at least one first stage multiplexer, with a single multiplexer being associated with each sub-group of signals, each multiplexer being operable to combine the up- or downconverted signals of each sub-group into a sub-group signal ;

0. The data transfer circuit as claimed in claim 9, in which the group signal is plural and transferred using coarse wavelength division multiplexing, thereby increasing scalability.

1. The data transfer circuit as claimed in claim 9 or claim 10, in which the individual converted input signals in the combined group signal are spread over a chosen bandwidth with a chosen channel spacing, such that neighbouring individual signals will not interfere with each other for a pre-defined amount of frequency shift.

2. The data transfer circuit as claimed in any of claims 3 to 1 1 inclusive, which includes a filter associated with each converter thereby to remove any additional frequency components and leave only the converted frequency component.

3. The data transfer circuit as claimed in any of the preceding claims, which includes a tuning, matching and detuning circuit, in no specific order, and a low noise amplifier associated with each input connection.

4. The data transfer circuit as claimed in any of the preceding claims, which includes a resonator tuned to an RF excitation pulse frequency of the MRI machine thereby to provide on-chip power generation.

5. The data transfer circuit as claimed in any of claims 1 to 13 inclusive, which includes a non-ferrous electrically conducting enclosure or cage, operable to shield electrical fields.

6. The data transfer circuit as claimed in claim 15, in which the enclosure defines an aperture aligned with the output optical connection thereby to align an optical fibre and the output optical connection, in use.

7. A data transfer circuit for an MRI machine having a plurality of receiver surface coils, the data transfer circuit including:

at least one combiner operable to combine the plural input signals received from the input connections into an electrical combined signal; and an on-die optical driver and light emitter operable to modulate the combined signal into an optical output signal, thereby to transmit the output signal optically to a remote data processing unit,

the circuit being an integrated single-die circuit.

8. The data transfer circuit as claimed in claim 17, in which:

the integrated circuit is a silicon circuit; and

the light emitter is an on-die silicon light emitter.

9. A data transfer method for an MRI machine having a plurality of receiver surface coils, the data transfer method including: receiving, by an integrated circuit, a plurality of electrical analogue input signals from the receiver surface coils;

modulating combined signal into an optical analogue output signal by means of a light emitter.

20. The method as claimed in claim 19, which is implemented on an integrated single- die silicon circuit and in which the light emitter is an on-die silicon light emitter.

21. The method as claimed in claim 19 or claim 20, which includes converting each received input signal to an intermediate signal having an intermediate frequency prior to combining. 22. The method as claimed in claim 21 , which includes plural stages of converting and combining.

23. A data transfer system for an MRI machine having a plurality of receiver surface coils, the data transfer system including:

at least one data transfer circuit as claimed in any of claims 1 to 1 6 inclusive;

The data transfer system as claimed in claim 23, which includes a plurality of data transfer circuits for large scalability. 25. The data transfer system as claimed in claim 24, in which the optical fibre from each circuit is grouped into a fibre bundle for optical transmission to a remote data processor including an optical receiver, separator and ADC.