Research And Design Of Digital MRI Spectrometer System

Magnetic resonance imaging (MRI) is an imaging technology which using the nuclear magnetic resonance (NMR) property of organism. It has many features such as high resolution, multi imaging parameters, non-invasive and arbitrary slice imaging, etc. Now MRI is one of the most important tools in medical diagnostics and widely used for in-vivo studies.As the essential part of MRI system, spectrometer controls the time sequence, transmitting, receiving and processing of signals. Its performance is one of the most important factors to affect image quality. The design of spectrometer is very complex, so there are only a few foreign companies which can provide commercial product, such as RI, SMIS, Medison, etc. At present, most spectrometers need to be imported in our country. Research and design of homebuilt spectrometer is very important to the development of MRI technology.In this thesis, we conducted thorough researches on MRI theories and spectrometer technology. A new digital spectrometer based on PXI (PCI extensions for Instrumentation) architecture is proposed, which is constructed with the general PXI platform and boards. Software algorithm is designed to calculate and process digital signals, and PXI boards are used to implement A/D, D/A conversion. Software and hardware are independent in this design. We firstly introduce the integrated design of MRI spectrometer, including module structure and work flow. Then the implements of all modules are presented. The calculation and process of digital signals are emphatically analyzed. During the process that gradient magnetic fields are produced, the rapid switches of gradient fields will cause eddy current, which results in artifacts in the image. We conducted researches on eddy current compensation, and appended compensation function in gradient generator of spectrometer. There are probably phase artifacts existing in MRI image. In order to improve the quality of image, we present a correction method in this thesis. Our meaningful and detailed research work is organized as follows:1) A new digital spectrometer based on PXI architecture is proposed. Firstly, time sequence is quantified according to the MRI scan sequence. Digital radio frequency (RF) pulse and gradient pulse are calculated at the same time. Then the digital signals are converted to analog signals by PXI boards. MR signal produced by imaging objects is digitized and processed with digital quadrature detection and filter. Finally, we reconstruct the image. Compared to traditional spectrometers, the software algorithm is independent of hardware. The main features of the proposed spectrometer are simple architecture and high performance. The special electric signals of PXI can be used to implement clock synchronization and accurate trigger.2) Design the RF transmitter and RF receiver of spectrometer. We focused on calculation of digital RF signal and processing of MR signal, including quadrature demodulation, decimation, filter, etc. Digital generation of RF pulse makes it easy to control the frequency, amplitude and phase of signal. Application of digital quadrature detection avoids the spectral distortion due to amplitude and phase errors between two channels of traditional detection method. A/D and D/A boards are used to convert signal, so we don't need to design hardware.3) We made studies on eddy current, and presented a method of eddy current compensation based on pre-emphasis of gradient current. Gradient field is measured with Faraday's induction theorem and data acquisition technology. We can acquire compensation parameters rapidly utilizing data fitting and feedback control method. The experiments prove that the method is efficient to improve the image quality.4) In the process that MRI signal is quadrature demodulated by RF receiver, the initial phases of carrier signal exist errors, which results in artifacts in the image. In this thesis, a method based on reference scan is proposed to reduce the artifacts due to demodulation phase errors. Firstly, reference scan without frequency encoding gradient and phase encoding gradient is executed before MRI pulse sequence, and an echo train is acquired. We can calculate the phase errors of each echo in the reference echo train. Then the compensation matrix is constructed and used to correct the image. Artifact can be removed with this method.