Authors: Felipe Pelizaro Gentil & Daniel Consalter (Fine Instrument Technology)

Magnetic resonance imaging (MRI) techniques are in constant development. Its understanding requires knowledge in classical physics and quantum mechanics. The first studies were done back in the 50s by two collaborators from the University of Harvard and Stanford: Edward Purcell and Felix Bloch. Only 15 years later, Jasper John was responsible for doing the first magnetic resonance signal. A few years later, Paul Lauterbur was capable of reconstructing the first image by sending signals through a variation of a magnetic field.

A simple MRI study follows the following steps:

  1. The sample is set inside a homogeneous magnetic field, making the hydrogen spins of the sample being aligned to the magnetic field and the protons of its atoms turning around its own longitudinal axis like a spinning top.
  2. When a radio frequency (RF) pulse is applied directly to the sample, the proton feels its field as a torque and changes the direction of its turning.
  3. As soon as the pulse is over, the proton spin tends to realign to the magnetic field. During its returning, an electromotive force is inducted to a coil located near to the sample.
  4. This signal is codified by the magnetic field gradient produced by gradient coils coupled to the sample.
  5. The coil receiver is connected to the digital spectrometer in which the signal is processed sent to the computer, responsible for recreating the image using a Fourier Transform.

The digital spectrometer, also known as console or control module, is an electronic device of great technological complexity whose main function is to control an MRI experiment in a way to produce, codify and acquire the signal from the experiment. Doing a musical analogy, if the MRI experiment was the song, the spectrometer would be the conductor, the pulses generated in the spectrometer the notes and the group of sorted pulses, the sequence of pulses, the score. The spectrometer’s main function is to control the experiment as a whole, as if it was the brain of the entire device. FIT’s digital spectrometer project evolves the electronic module and the software development which works as an interface for the user to program and compile sequence of pulses besides controlling the parameters of the experiment. A new programming language was also designed and developed to help developping sequences of pulses.

Why FPGA? At the beginning of the development, there were a few options to be the base of the entire project such as microcontrollers, DSPs (Digital Signal Processor) and FPGA (Field Programmable Gate Arrays). For the following reasons, the choice was quite easy.

  • FPGA is a generic component, making it possible to become any of the other options, like the DSPs or the microcontrollers.
  • The interconnections between the FPGA modules is programmable by the compilation of a code which describes the hardware. In this way, time precision is guaranteed.
  • There are FPGAs available in the market that filled perfectly our needs with all the peripheral modules.

The architecture is similar to any other signal processor equipment plus the wave generators. Every system is temporized by a high precision clock as shown in the picture above.

Spectrometer Main Diagram

The project was designed to use a Master/Slave architecture. In other words it means that one device (master), or in this case module, controls one other module (slave). The flow of control is always from the master to the slave.

Spectrometer - FPGA Diagram

The Receiver is responsible for receiving the signal captured by the coils, demodulate this signal and filter. This module is connected directly to the FIT Bus and the TopNet (responsible for the communication and will be shortly discussed further).

The TopNet module is the one that makes the communication directly with the software. It implements an UDP protocol and is responsible for packing all the data that has to be sent.

The Gradient is responsible to generate a single point every 1us, which means 100 cycles of the clock. This point represents the three space axes (X, Y and Z) plus the magnet field (B0). The gradient is also responsible to generate the signal that will be sent to the gradient coils, responsible to generate their own field and codify the signal.

The RF is the module that generates the radiofrequency signal, which is sent to the sample through a RF coil and consequently changes the direction of the spins, starting to produce the resonance signal.

The picture bellow was created with this spectrometer in a Centaury3000 MRI system from XinalMDT.

image created using spectrometer in a Centaury3000 MRI system from XinalMDT

Our results validate the use of a FPGA not only for medical applications but also for chemistry, oil and agriculture. Its use is applied for samples with even number spins (H1, C13, F15, etc.) and magnetic field with resonance between 10 MHz and 50 MHz.