06-07 The Complete Imaging Experiment

In two-dimensional MR imaging, a slice is excited, for example, by applying a selective RF-pulse in the presence of the z-gradient. There are two different methods of obtaining spatial information for the remaining two dimensions:

06-07-01 Frequency-Encoding Only

The remaining two gradients (x and y) are now combined with a resulting gra­dient of a certain strength and spatial direction. The FID is recorded in the pre­sen­ce of this gradient. Then the combined gradient is rotated a certain angle. Once again the FID is recorded, and this process continues until enough in­for­ma­tion is collected.

Based on the frequency spectra, an image can be reconstructed using the back­pro­jec­tion method, as depicted in Figure 06-13.

06-07-02 Two-dimensional FT Method

This method is a combination of phase- and frequency-encoding. Today it is the standard image formation method.

One gradient, for example the y-gra­dient, is switched on and the spins are al­lowed to dephase in the presence of this gradient. After a certain time, the y-gra­dient is switched off and the FID, or alternatively the spin echo, is recorded in the presence of the x-gradient. Thus, the y-gradient serves as the phase-en­co­ding gradient (often called preparation gradient) and the x-gradient is encoding the frequency information (readout gradient). The system is re-excited, but with either the duration or, more commonly, the strength of the y-gradient changed.

The whole process is repeated n times for a resolution of n pixels in the y-di­rec­tion, with a different phase-encoding gradient applied for each excitation.

Since the field inhomogeneity effects will be the same for each repetition, they have no effect upon the final image; rather, they constitute a baseline. This is one of the main advantages of the 2D Fourier-transform technique. The re­sul­ting 2D raw data matrix is processed using a 2D Fourier-transform to produce a 2D image (Figure 06-20).

The combination of frequency- and amplitude-stepped phase-encoding tech­ni­ques is called 2D spin warp imaging [⇒ Edelstein].

Figure 06-20:
2DFT owes more to MR spectroscopy than to CT reconstruction algorithms because both am­pli­tu­de and phase information are acquired to spatially encode the signal.

The first step generates a one-dimensional projection, but here a phase-encoding gradient is ap­plied just before the original gradient is turned on. The phase-encoding gradient is applied in right angles to the original gradient and its duration or am­pli­tu­de is successively increased.

The corresponding points from each projection are Fourier-transformed a second time to ge­ne­ra­te the final image [⇒ Pykett].

Figure 06-21 summarizes an entire 2DFT imaging experiment, in this case using a spin-echo sequence, although one can use nearly any other pulse se­quen­ce applied in clinical MR imaging.

Figure 06-21:
Complete 2DFT spin-echo imaging experiment. The procedure consists of the selection of the 90° and 180° RF pulses, a transverse slice through a brain (z-direction), phase-encoding (y-direction), and frequency-encoding (x-direction).

The phase-encoding gradients change the phase in the respective row of the transverse slice; the frequency-encoding gradients allocate a specific frequency to each column. Combining both phase and frequency information allows the creation of a grid in which each pixel possesses a distinct combination of phase and frequency codes.

The entire procedure is usually repeated 256 times, with changing phase-encoding gradients to produce a 256 × 256 image.