Okay. Let's talk about Echo Planar Imaging. As mentioned in the previous video lecture, we talked about the fast spin echo, which uses multiple 180 degree RF pulses to acquire data multiple K-space lines within one after 190 degree RF excitation. So, similar thing can be performed for the gradient echo imaging tool. So, in case of spin echo, 180 degree RF pulses generate an echo. So, in case of gradient echo, gradient prefetching and reversing readout gradients generates gradient echo. So, to generate multiple gradient echoes, we have to switch readout gradient multiple times, then we can acquire multiple gradient echos. So, let's try to compare these two pulse sequence diagrams. So, one is the basic gradient echo pulse sequence diagram, which is just same as the gradient echo sequence that we just mentioned. Then each readout gradient will fill one K-space line, and then we apply next RF excitation and then acquire another phase encoding line. So, each K-space line is acquired after RF excitation. But in case of Echo Planar Imaging, so which is our first occasion mode in the gradient echo position. Then after activation of RF pulse, and then we can read echo, the first echo up to this point which it is the same as the conventional gradient echo. But just we can apply for another gradient in the opposite direction, in the same duration and strength as the first readout gradient, and then both the polarities are plotted. Then, this gradient will accelerate Taylor decay, and then it will be reversed in the middle of this gradient again, so that then we will observe a gradient echo again. Then we can apply for positive readout gradient two, and then we can acquire dephased and leephased echo, and this procedure can be repeated. This signal decay will follow T2 star. T2 star signal decay is maintained, but we can switch readout gradient multiple times to oscillate decay faster, and then recover back to T2 star, and oscillate faster and recover back to T2 star, and this procedure is repeated. This procedure is very similar to the application of multiple 180 RF pulses. But the signal is not going to recover to T2, but it will be recovered only to T2 star in case of gradient echo. Okay, this sequence is called Echo Planar Imaging, EPI. So, this Echo Planar Imaging is the fastest MR imaging sequence, and multiple echos are acquired from alternating frequency including gradient as shown here. Then phase encoding is just used as a blip between the readout pulses, and phase encoding blips to fill the whole K-space region. We will talk about them in the next slide. Then acquiring a whole imaging typically, we didn't want RF excitation. Why RF excitation? We can fill whole K-space line for one slice. How can it be done? We will talk about that here. So, in case of just conventional gradient echo, we apply for phase encoding gradient and readout pre-phase gradient, and that moves the K-space location to the lower left corner, and then readout gradient will read data all the way to the right as shown here. Then next excitation will start from the K-space center again, then move here, and then fill next K-space line. This procedure is repeated with different RF excitations in case of gradient echo imaging. But in case of Echo Planar Imaging, so up to this point, and this point, for the first tackle it is the same, and then we apply for a phase encoding blip, and it will move one step up. Then next readout gradient with opposite polarity will move K-space location from right to the all way left. Then phase encoding blip will move increase the step, one step for all, and then now the positive readout gradient is applied and then that will acquire data from left to the all the way right, and then this procedure is repeated after excitation. So, both readout and phase encoding could pre-phase gradients move the K-space position from the center to the lower left corner which is common for the conventional gradient echo and the API, but alternating frequency encoding gradients fill the data from left to right and then from right to left repeatedly. But phase encoding blips ship the K-space location step-by-step as shown here. So, that is the mechanism of filling the whole K-space after one RF excitation, so that is the concept of Echo Planar Imaging. This image is an example of image from Echo Planar Imaging. In the viewpoint of our pulse sequences there are two different ways of achieving Echo Planar Imaging. One is gradient-echo based Echo Planar Imaging, and the other is spin-echo based echo-planar imaging. So, for the Echo Planar imaging, we can keep switching the readout gradients with different alternating polarities. So, we just talked about the pulse sequence for the gradient echo EPI, so slice selection, refocusing of phase encoding gradient and readout pre-phase gradient, and then readout gradient is kept switching. Then these phase encoding blips will have opposite polarity compared to the initial phase encoding gradient. Then we can acquire multiple kth echoes, which you will fill multiple K-space line. So, this is sequence diagram for the Gradient Echo-EPI. In case of Spin Echo -EPI, it's just slightly different. There will be 180 degree RF pulse, like a spin echo. Then in this case, readout pre-phase gradient and readout phase encoding gradient, polarity should be changed compared to the gradient echo case, because this 180 degree pulse will change the polarity for the whole spins. So, this polarity should be changed, which is same as conventional Spin Echo Imaging case. Then we can apply for readout switching with polarity. Opposite polarities. They will keep repeating as shown here. Then, the echoes filling the K-space center will be formed from the 90 degree to 180 degree and 180 degree to the echo center. So, this center-echo will fill the k-space center and then, that time point is maintained for the time of Spin-echo. Spin-echo will be formed from this 90 degree and 180 degree added pulse. This time is half of the echo-time and then, that is going to be half of the echo-time and then, this lead out it adjusted to have the echo filling the k-space center will correspond to the echo-time of this Spin-echo, that is the Spin-echo EPI. So, in terms of pulse sequencing two the only difference between Gradient-echo EPI and Spin-echo EPI is the existence of 180 degree RF pulse and the change in polarities of readout and phase encoding prephase gradient as shown here. Let's try to compare Gradient-echo EPI and Spin-echo EPI. So, in terms of flip angle, any flip angle can be used for creating a Gradient-echo EPI. Typically, you can apply for Ernst angle as a flip angle, but in case of Spin-echo, we typically use 90 degree and 180 degree. And K-space center echo is still gradient echo, but the k-space center should be Spin-echo, as I mentioned in the previous slide. And readout is going to be gradient echo and then, scan time is both of them have quite short. And compensation of field inhomogeneity is not possible for the Gradient-echo imaging, but a Spin-echo imaging is yes, and also this Spin-echo EPI cases this is center-echo that fills k-space center is T2-weighted. In case of Gradient-echo its is T2*-weighted, so that is the difference though there is some field inhomogeneity compensation for the Spin-echo EPI, but there'll be no field inhomogeneity compensation for the Gradient-echo API. Let me talk about the segmented EPI which is called multi-shot EPI. So, multi-shot means, each shot means one RF excitation can be used to fill the whole k-space and then, one-shot means one RF excitation. One shot EPI means whole k-space is filled after one RF excitation, but that is possible, but you can acquire one case k-space line from multiple-shots and that is called segmented API. So, segmented EPI utilizes multiple RF excitations to fill one whole k-space compared to the single-shot EPI which utilize just one RF excitation to fill the whole k-space. And this multi-shot EPI provides higher antenna and also better image qualities and less distortion, but it requires a little bit longer scan time. So, this is pulse sequence diagram for the multi-shot and the sequence diagram is almost the same as the single-shot case. But, this whole multiple-shots, in this case four shots, fill the whole k-space, so each shot will fill k-space like this. One line is filled like that and then, it will jump four steps up and then, move to fill the next line and then, four steps up and then, fill that lines, and that will be repeated for the one-shot and the next shot will fill the second phase-encoding line. Start from second phase-encoding line and then, jump four steps, and this procedure is repeated, and third one starts from the third phase-encoding line and jumps in the same four steps. The fourth step is also same thing, so that will cover the whole k-spaces. So, total phase-encoding lines divided into multiple segments and the number of lines per segment, is going to be total phase-encoding line divided by number of segments, so that is the famous echo train length for the fast spin echo. Because EPI is accelerated there this is great advantage, but it caused a lot of problems, and one of them is ghost artifacts and the reason is we fill k-space along with positive gradient and then, negative gradient and positive gradient and negative gradient. If gradient systems are perfect, these two echoes should be aligned exactly in the middle, but that is impossible in any gradient. MRI system have some imperfection in the gradient, so there is almost no exception. So, these EPI sequence has positive even echoes and odd echoes have different locations on the k-spaces. So, ghost artifacts can be induced in EPI because of gradient imperfections. Some ethical and some other gradient imperfections that cause mismatch between phase-encoding lines from all the echoes and even echoes as shown in this figure. And then, ghost will look like this, there's main object and then, some object exist here and here. So, if there are N pixels across field of view, as shown here, and this aliased ghost appears at the position shifted by N/2 pixels relative to the original location as shown in this figure, so because of that this artifact is called N/2 ghost artifact. So, the artifact can be suppressed by phase map acquired with the same EPI scan, but without phase encoding gradient or blips, and this scan is called a reference scan, this equal to reference scan. And the phase information from the reference scan can be used to compensate for this value induced by this mismatch shipped, they can be used to minimize ghost artifact. So, these are the advantages and disadvantages of EPI. The greatest advantage of EPI is the fastest. This sequence is the fastest MR imaging sequence. And the whole brain images can be acquired in a few seconds what do they even one second or even below one second is possible by combining with some special acceleration techniques. And it's applicable to whole functional MRI or physiological MRI like perfusion imaging or diffusion imaging. And disadvantages of this EPI is severe distortions and signal drop out and some ghost artifacts and also spatial resolution is low because the signal-to-ratio is low. These are the advantages and disadvantages of EPI.
