JEMRIS  2.8.1
open-source MRI simulations
First steps tutorial

In this tutorial the user is introduced guided through the setup and performance of an EPI experiment. This step by step introduction shall enable the user to setup own experiments thereafter.

To proceed the user must have successfully compiled JEMRIS 2 as described in the ug_intro_installation section.

Setup the EPI sequence

We will now start to model an EPI sequence with the following protocol: TE 50ms, TR 100ms, FA 90, Imaging matrix 64x64, Maxmum gradient amplitude 40 mT/m, Maximum slewrate 200 mT/m, and FOV 100x100.

You will now see the minimum requirements for a sequence. The top node (outer most loop) of the sequence, C1, short ConcatSequence1, and the paramters, P. Let us start with the parameters. The parameters hold information valid for the whole sequence. These include in order from top left to bottom right: name, FOV in x, y, and z direction,maximum gradient amplitude, maximum slewrate, TD, TE, TI, and TR.

Let us now model the sequence.

Since we do not loop over more than one slices you do not need to change any of the parameters of Sli.

We want to start with inserting the excitation puls. Pulses, which are played out simultaneously are contained in atomic sequences.

An atomic sequence is appended to Sli by the name of A1.

Since we do not care for slice selection for this tutorial, we now insert an hard RF pulse.

When you hover over the fields a short description is shown for every module.

Your sequence should, so far, look like this:


The excitation is now sufficiently defined. We now need to insert into the sequence the dephasers in read as well as PE directions. For this purpose the next atom needs to be added after the Exc atom.

The areas for the dephasing gradients can not be set yet, since we have to prepare the actual readout and phase encoding gradients first. But first we need to introduct a dead time to match the center of the EPI readout with TE from the parameters.

You will find details on delays here: ug_usage_sequence_gui_intro_seqmod. Once again the delay cannot be complete entirely besides that we know that TE starts at the centre of the excitation and ends at the centre of the EPI readout.

One could now type in the value for TE in Duration. But a very convenient feature of JEMRIS helps us minimize the effort for future changes to our sequence such as changes to TE. Namely that every module is able to observe values in other modules of the sequence; may they be dynamically changing over the duration of the sequence or static.

We would like to observe TE from P. We inform DTE in the following manner:

You will find a short introduction to the concept of observing attibutes here: ug_usage_sequence_gui_ginac_obs. Now we can use this new attribute known to the module as a1.

You should be looking at a sequence representation looking as follows:


Let us now insert and compose the EPI readout. For this, we would like to implement a loop structure for one readout line and one phase encoding blip to run through all phase encoding steps.

We have to define the FlatTopArea and the number of sampling points on the flat top of ROG. The latter is rather straightforward. We need to observe the imaging matrix size defined in P.

This is the first attribute and known to ROG, according to above, as a1.

For defining the flat top of ROG we have to know the maximum k-space vector in readout direction and the number of the line, which is acquired. We introduce these two attributes to Observe. Observed values are seperated by / (slash).

ROG now knows the two new attributes as a2 an a3. The FlatTopArea then should change the polarity for every second line and go from -KMAXx to +KMAXx.

Here the integrated library for symbolic mathematics, GiNAC helps us to calculate the FlatTopArea by the evaluation of a symbolic formula.

Let us define the Area of the PE blips, PEG. This is a very simple assignement. The area will be the size of the phase encoding steps in Ky direction.

The following image reflects the content of the sequence GUI up to this point:


Now that ROG and PEG are defined we can go back and define the missing parameters of RDp and PDp. RDp should send the magnetisation to -KMAXx and PDp to the top edge of the K-space (KMAXy).

Let us now define the missing parameters of the delay DTE.

The only missing element is now the constraints associated with TR. So will add a delay to the end of the sequence to suffice those constraints.

We are done now with the EPI sequence.

Let us, finally, have a look at the completed single slice EPI sequence, the associated sequence diagram, and the k-space trajectory reflecting the data acquisition scheme in the following three images:


Click now on the check box Sequence Diagram at the top to see the currents played out during this sequence.


Click now on the check boxes k-space trajecory and continuous to verify the sampling pattern and time evolution from early to late.


Let us go on to simulate the sequence: First steps tutorial - simulation

-- last change 17.06.2016 | Tony Stoecker | Imprint --