Introduction to DMIWizard

January 21, 2020

ID4769

To CiteDCA Citation Guide

00:00Hello and welcome to a brief introduction video to D my wizard and specifically a version 1.1.
00:08D my wizard is a matlab based GUI aimed at the processing of 1D and 3D deuterium metabolic imaging data acquired on a number of.
00:19Of platforms. This video just gives you a brief walk through of the most common operation that you do within their software.
00:29If you have any kind of trouble or any kind of questions or require modifications.
00:35Please contact me directly at aerobicgraphyield.edu.
00:41OK, at this point, I'm assuming that you have downloaded the my wizard software from the website.
00:47In this case. It is D.
00:48My wizard version 1.1. Zip and you can now go ahead an unzip that file.
00:53And normally I'd like to copy this and put it in a place where I like it to be in this case.
00:58Let's make a dmi directory on the C drive it will paste it there.
01:04And and ideally you've also downloaded the dmi wizard data file.
01:09It also zip file that contains a number of Emerson.
01:13MRI files that you can use to practice.
01:16The Dmi Wizard. So far with it also zip file so go ahead and unzip that or copy that and put it again in the in the place where you
01:26want it to be.
01:29There we go. So the first thing you typically want to do is going through the demon Wizards Directory and then go through the D my wizard settings dot text
01:40file. At name implies, it is a simple text file it holds some information about the colour scheme of the Demon Wizard.
01:49GUI's about the font name and sizes and also about the default part name.
01:54So you want to change this to the directory name that you have on your computer in this case.
02:00It's the C drive and then the dmi directory that we just made if you're happy with you can save that.
02:08And you can see dash already showing premade settings as well.
02:13That allow you to do a dark mode.
02:15GUI setting and you can change those as you see fit.
02:20OK, so normally then I would like to copy the directory name and paste it into Matlab.
02:28And then the GUI can be started by just typing D my Wizards in the command line.
02:35And then you gonna get your main window as shown here.
02:39The main window has basically 2 main menus.
02:43One is for the processing of one dimensional spectroscopy data and the other one is for the processing of 3,
02:50Dimensional Amar Spectroscopic Imaging data.
02:54So let's first quickly go truly one DM arrest data.
02:58And used to emphasize that this video is not an exhaustive manual to all aspects of their software that you can find in the menu.
03:06As I'll show later this is just a quick walk through on some of the most important features.
03:12So this is the menu for the one dimensional spectroscopy.
03:16The lines in green and boxes in green and typically related to the file loading and file handling items things in blue are spectral processing things in yellow are related
03:29to baseline correction integration and things in red are associated with a chemical shift referencing so the first thing you typically want to do,
03:39if you want to load in a data file.
03:43So that's where the D my wizard underscore data comes in.
03:47There's basically 4 different types of files there is a 3 dimensional.
03:52Dmi there is a one dimensional spectroscopy.
03:55But one average. There's a one dimensional spectroscopy with 180 averages,
04:00and there is 3. Dimensional MRI in this menu were interested in #2 and #3.
04:05So you can open that and then you can read this being brokered data you can read the FID file.
04:12At this point, it hasn't loaded the data.
04:15Yet yet just selected the file that you want to load you wanted,
04:19and indicate which vendor the data is from in this case,
04:22it is broker and then you can go ahead and load the data.
04:26And then the data will be displayed here in Figure 1 the time domain data.
04:31The red and blue on the real and imaginary aphids.
04:36Ann.
04:37In this case, these boxes here they indicate the data structure and Emmanuel have a lot of information about this.
04:44But basically here you have to indicate how the data structure is at the moment.
04:49It is just 1F ID so all of these values are one so you can go ahead and do appreciate the information.
04:56And there, you basically have your spectral or frequency domain over there.
05:02Ann. You can see it's not properly centered so you can hit this zoom out button and then you can go ahead and zoom it in.
05:11Closer to a manual zoom again,
05:14I can type in numbers here minus 0.4.
05:19Mine is your .5 for example,
05:21loops.
05:23Now he can type in one.
05:27How to do that plus or minus buttons to to achieve the zoom that you want so the first two are zontal zoom.
05:35These 2 are a vertical's room.
05:38And then the one on the bottom is a vertical school zoom with a factor of 10 in intensity to every time you go up is a factor of 10,
05:47so you just make it as you see fit.
05:50New bit of horizontal zooming etc,
05:52etc. And you can always zoom out,
05:55then you go to the maximum spectral width and you can of course,
06:00zoom in again.
06:02And indeed 2 boundaries are set based on the cursor positions.
06:07An traditional spectral parameters are you can do something from line broadening your phrases 3 hers.
06:12Gaussian requires an additional forget transformation and you can see that the peaks become a bit broader.
06:19Of course, his spectrum isn't properly in phase so you need to face corrected.
06:24This is zero order phase and there's also a first order phase as well.
06:30And then you can also do baseline Correction,
06:33an integration. I'm going to not talk about that.
06:35Today I will talk about briefly about the frequency referencing so if you click this button.
06:41You can pick you can pick the frequency the pic that you want.
06:46And then if you know the chemical shift and I think this will be water.
06:50Then we calibrated and now the access isn't in PM so you can zoom out again and then zoom back in.
06:58And you can see. Now you have your beaks at the chemical shift in PM.
07:03Can have a quick look at the other file as well?
07:06Where we have 180 averages?
07:09If you load in the raw data that job zero file.
07:14You will get a lot of data you actually get 180 FI DS because they're not added together.
07:20Yet so if you want to see all the individual ones you have to indicate that the number of initial Fridays is other than 80.
07:32Feed and reload it. You know basically show you the first F ID and if you now do a free a transformation you now have.
07:41180 Spectra, he will display the full spectrum,
07:44but you can go to the second spectrum that dirt spectrum.
07:48All the way to the very end and so this allows you to see how stable your data is if there's any kind of artifacts or frequency drifts in the
07:57data and they can also be undone with these menus over here.
08:04OK, so that's where all I want to say about the one dimensional data processing.
08:09There's a lot more information in the manual the manual can be found in the Dmi Wizard Directory and it is right over here.
08:17D my wizard version 1.1 manual.
08:19And it has a lot of information about the spectroscopy menu and also about the 3 dimensional.
08:27MSI menu and we'll talk about that right now,
08:31so. We can go to that menu.
08:35Is a fairly similar appearance again things with blue are spectral parameters things in green are associated with file loading things with red in this case are amaray related things
08:48and things were yellow. Our metabolic mapping parameters,
08:52so let's do the first thing is,
08:54we select the dmi data.
08:57Can go to the data directory?
08:58Now it is file number one directory number one that you're interested in and this being broker data.
09:04You need to load the raw data dot job 0.
09:07Again, this only selects it's loading it.
09:11Gives you some parameters in the matlab window about field.
09:15A few number of encodings,
09:17etc, etc. And then it shows you one of the Fridays in this window.
09:23Can you have your information again after selecting for example,
09:27inappropriate waiting you can do a free a transformation?
09:31Take of course, a few seconds because it's a very large data set and then here is the spectrum from disposition disposition.
09:40One eight one out of a data set 2D2.
09:42By 2 by 2 by 16.
09:43So it may be completely wrong location and that's why the spectrum doesn't look very interesting.
09:50So of course I deal with that is,
09:53you want to load in an associated MRI file that can show you where the voxel location is so go ahead and go to the data directory and now it
10:03is directory #4, the MRI.
10:05He loaded F ID file.
10:08Launch.
10:10And then if we ate information and you can see how the Phantom looks like it's a number of tubes with different chemicals in each of the tube so at
10:20the data set of the MRI is 96.
10:22By 96 by 64, so let's go to the centre slides right now,
10:26which is 2D2. And then will also go to the center location in a dmi data set which is a 2D2.
10:34By 2 by 2 by 16,
10:36so let's go to Z slice #8 and there,
10:39you can see the Red Square.
10:41That indicates the voxel location at position one in X.
10:478 in Y an 8 Ng.
10:48It is the true plane dimension,
10:51so if you move the red volume to a more interesting location.
11:02There we have a spectrum and so he began zoom in similar to the one dimensional menu.
11:08So that we can see all of the peaks and we can walk through the data.
11:13And. As such. Right they have different chemicals in different locations.
11:21OK, we can also display the data.
11:24By hitting this Emirates I button there.
11:28Will take a few seconds because it's it's quite a large data set to the 2 actually 2 bye?
11:341024 data point I believe.
11:37With after a few seconds,
11:39he will show you, your 2 dimensional grids on this particular Z slides and you can see you have signal where the tubes are and there's basically no signal where
11:49there are no tubes. Now there are certain spectral that are in red an spectrum that are not in red there in blue now that difference is if a particular
11:59virtual location. The signal in that location is it above the threshold or not.
12:05In this case, these are in red there above the threshold,
12:09the ones in blue or not so by lowering the threshold.
12:13Let's say to 5%. We can change that and if we now show the spectroscopic image again.
12:20Then you will see that you have more red Spectra.
12:28There we go, she's now pretty much everything that we want everything that we want to have signal has signal now?
12:36Why is it important that is important because some operations like linear prediction.
12:43They only work on spectral that are in red and that his basement done just to save some time.
12:50There will be an awful lot of data that have no signal and it will be wasteful to do complicated math on.
12:58Virtual location that have no signal.
13:01Charlottes quickly talk about linear prediction.
13:04As you can see if you load it in the dmi data.
13:10That if it is. Typically,
13:14the way they did my data is acquired it is basically a pulse acquire sequence with the phase encoding gradient in it.
13:22The presence of the phase encoding gradients basically delaye simple acquisition by a little bit,
13:29for this particular scene because it is 0.6 milliseconds.
13:33The spectral width is 2 1/2.
13:35KH making the dwell time,
13:37400 microseconds. So there is about a one and a half data point of delaye between the excitation.
13:44And the first point of acquisition,
13:46let's make that tool. There's two data points that are missing now linear prediction allows you to back calculate those 2 missing data points.
13:56So if you type in here 2 and hit linear prediction.
13:59You basically have to indicate the part of the spectrum that contained signal.
14:05And then you can see here it will do the calculation for 2500.
14:09Spectra that's already a lot of Spectra,
14:11but it goes fairly fairly fast.
14:13But if you didn't have to threshold if we would do it for all Spectra.
14:18Which would be 16,000 so you've definitely improved on the speed?
14:25So if you give it a second you will see that all of this process is done.
14:29All of the signals will be in face never we up and positive.
14:33At the moment, there, not as you can see because of that 2 points data gap.
14:40Cable almost done here.
14:52And there, we are so now we have to display it.
14:57There we are so now if we now work,
15:00a little bit through the data.
15:02You can see that pretty much the picture up M positive if we go down a little bit here.
15:07Go to a different volume see all of the peaks are up and positive.
15:11Now there is important for the final step that I want to talk about which is the parameter mapping parameter mapping allows you to make an image of a given
15:20peak so for example, if we click here parameter map and we select the center peak right over there.
15:27He basically generate an image it integrates this part of the spectrum between the 2 lines and then displays,
15:35it at and as an image but that is a phase sensitive operation.
15:39The signals have to be up in positive if the sequence are out of phase you may end up with zero integral.
15:47You got to do that for other peaks as well,
15:49and you will see you get in a different spatial distribution because the chemical content of these different tubes is different.
15:58OK let's go back to the 3rd Speaker looked a little more interesting.
16:04So we can also do is you can overlay this on an MRI and MRI is really loaded so if we click this button over here.
16:14And we do the parameter net again,
16:16then you can see that you can overlay it with an MRI image.
16:20And so you can see,
16:21there are images now in the background.
16:24But maybe it's not bright enough So what we can do is we can make the transparency.
16:28A little bit lower on the metabolic image and so you can see the MRI a little bit better.
16:34And so there gonna have it,
16:35he MRI in the Emirates Ioannou overlapping with each other.
16:41We can also do an interpolation which that is more like a visual trick,
16:47but it can sometimes look look more appealing.
16:51If you will then you can have you have 3 types of images one?
16:57Is the native Amazon resolution.
17:00One is an interpolation based on FT and one is based on like convolution convolution typically gives the best results and the.
17:11This number 3 basically tells you how much smoothening has been done.
17:21Final thing I would like to remark is sometimes you're not interested in all the signals.
17:27So you can draw an ROI as well.
17:30And so let's say that we drawn ROI over here.
17:38Headsets if we now do the parameter map again.
17:44You will see that. Now he only overlays a map multiplied with the ROI answer.
17:50This could be a good option.
17:52If you want to have a brain only ROI for example.
17:56Now this is an early version of the email is it in the next version.
18:02And is mapping these mapping options will be a lot more sophisticated way you can do multiplication of Maps and a bunch of other of other things as well.
18:13So I encourage you to check the website frequently for newer versions of Dmi Wizard again.
18:20This is only a quick overview.
18:22If you want to have more details you can find it in the manual.
18:27But this is all I wanted to say about it.
18:30Thank you for using the my widget and let me know if you encounter any problems.