Updated June 2025
¶ PRIDE jobs: General information
The PRIDE interface (Philips Research Imaging Development Environment) is Philips' way of making it possible to process data on the scanner using custom tools. In the past, we mainly used MrCodeTools for this, but as that is no longer supported we now use COMPASS by Tesla Dynamic Coils for many processing jobs (B0 shimming, MP2RAGE, RF shimming, …). For more information on COMPASS, see here. In addition to COMPASS, there are also some ‘local’ tools available on the scanner. This page will give a quick guide to the PRIDE jobs on our system.
This page contains information about various different tools - please note that you can use the Table of Contents (on the left-hand side of the Wiki) to find what you're looking for!
¶ Getting started
As PRIDE jobs process acquired data, you first have to acquire data. After the acquisition and reconstruction have finished, in the sequence list on the right-hand side of the scanner interface, right-click the sequence you want to run a PRIDE job on, and select “PRIDE - Research”. This will open a window with a drop down menu, where you can select which tool you want to run: see the figure below.
Alternatively, PRIDE jobs can be attached to sequences in an ExamCard. If this is the case, they'll execute automatically when the sequence they're attached to finishes. For example, in the ‘Default scans’ ExamCard (under Hospital → Standard), all the drop-down icons (marked by orange arrows) indicate that a PRIDE job is attached to the various scans. By clicking the drop-down icons, you can check which PRIDE jobs these are: B0 shimming, MP2RAGE processing, data export, clearing the PRIDE cache, etc.
¶ COMPASS PRIDE jobs
We use the Compass software for various jobs processing images, such as B0 shimming, MP2RAGE processing, and B1 map calculation + RF shimming.
Make sure the Compass framework has been started up before any COMPASS job starts. In general, it's good to just get into the habit of starting Compass at the start of every scanning session (do, however, make sure to not open multiple instances of Compass if previous operators failed to close it at the end). If you forget, simply re-run the job by adding it to the B0map scan again by ‘dropping’ it onto the scan and removing the wait sign (no need to reacquire the B0). You can open Compass using a shortcut on the desktop, or by simply searching for ‘compass.exe’.
¶ Compass B0 shimming
Compass allows for B0 shimming using various types of ROI masks: Brain extraction, hand-drawn masks, thresholding, secondary masks, etc. As the vast majority of scans use BET brain extraction, this is the default setting and this will be described here.
After a B0 shimming job starts (either through a PRIDE job attached to a ‘B0 unshimmed’ sequence, or through a manual call), two windows should appear within a few seconds - see below. These show the histograms of the shimmed and unshimmed B0 values within the ROI (top figure), and the corresponding maps (bottom figure).
This generally works quite reliably. However, if you have any doubts about the performance or reliability (e.g. because there were crashes at the start of the session), there are a few things you can check:
- Subject name: At the top row, the name of the subject ('sub09' in the example) and the time/date when you started the session should correspond to your current session. If this is not the case, the displayed shim was calculated for an old dataset.
- Check that the histogram of B0 values becomes narrower after shimming and that the shim values in your scan are updated.
- Check that the mask is (visually) correct. In the figure above, the ‘Mask mode’ is set to ‘Trim’, showing the ROI only. The ‘Shim mode’ button makes it possible to switch between the acquired unshimmed maps and the predicted shimmed maps.
If the phase unwrapping of the B0 map is not complete, then the tool will not be able to find a good shim solution. You may notice the remaining phase wrap in the popup images, or you may simply have a terrible EPI image as a result. If this happens, you can do two things:
- Re-calculate the shim using Compass using a different FOV (start Compass.exe, choose tools, choose B0-mapping, add a manual mask excluding the offending region).
- Re-run the B0-map with a smaller value for the deltaTE parameter. 0.8ms instead of 1 is usually enough.
- Use the PRIDE job which uses the Universal shim as a starting point (see below).
For more information about B0 shimming with Compass, see this manual: /b0_shimming_instructions.pdf
¶ Compass MP2RAGE
In MP2RAGE acquisitions, gradient-echo images are acquired at two different inversion times. From these two images, a T1-weighted ("UNI") image and a T1 map can be reconstructed. The standard scanner interface does not facilitate this reconstruction, so we use a PRIDE-based Compass job to perform this reconstruction.
As can be seen in the figure in the ‘Getting started’ section above, we have set up Compass-based MP2RAGE reconstruction using two different masking strategies: threshold-based and BET-based (brain extraction). Threshold-based masking has the advantage that anatomical features surrounding the brain are maintained after reconstruction, but with the disadvantage that parts of the brain can be masked out as well in case of signal voids due to (mostly) B1+ effects. Other than this masking step, both tools work the same way.
Depending on the reconstructed resolution (and FOV) of the acquired data, MP2RAGE reconstruction jobs typically take somewhere between 1 and 5 minutes - so you might have to be patient, especially for resolutions well sub-1mm. Make sure that the reconstruction has finished before exporting your data, as otherwise the T1w and UNI images will not be exported. There are three ways to check whether the job has (successfully) finished:
- In the protocol list, the MP2RAGE job turns light blue (equivalent to finished scans) and gets a tick-mark;
- In the terminal-view of the PRIDE job, the final line will state that it has completed;
- In the ‘Thumbnail view’, a “WIP” scan has appeared immediately following the acquired MP2RAGE data - see below, where the green circle indicates the buttons where you can toggle thumbnail view (left = protocol list, right = thumbnail view), the blue arrow indicates the acquired data, and the orange arrow indicates the reconstructed data.
Through this thumbnail view, it's also possible to view and inspect the reconstructed UNI/T1w images within the scanner environment.
¶ Compass RF Shimming
The easiest method for multi-channel B1 mapping and RF shimming is to use the ‘FFE dyn’ (channel-wise relative B1 maps) and ‘B1 quad’ (quadrature mode B1 DREAM quantitative map) sequences in the default sequences, which are also visible in the second figure under ‘Getting started’ at the beginning of this page. Those both contain a PRIDE job for data export, which exports the reconstructed data for both acquisitions to pride folder. From there, you can load the data into the Compass GUI for B1 mapping and RF shimming.
<TBA: screenshots, settings, and output>
¶ Other Compass tools
Besides the Compass tools mentioned above, there's a lot of other options or modifications which can be made through the Compass GUI. These have not necessarily been tried-and-tested on our systems, but you're more than welcome to give them a go and explore whether they can be of use. In some cases, more information can be found in the Compass Manual (/compass_framework_manual.pdf). Please do not modify any existing Compass configuration files without consulting Spinoza staff.
¶ Custom PRIDE jobs
In addition to Compass jobs, there are some custom (home-built) jobs which can be used for some specific applications. These are often made for a specific application. A few examples are described below.
¶ Universal B0 shim
No two people are the same, but most people are quite similar: they have a head with a brain, a skull, and some fatty tissue around it, with a body attached at the bottom. This means that the B0 attenuation caused by the body is very similar between different individuals, and that a correction which works for 10 people will probably also work reasonably well for and 11th individual. Based on this, we've designed ‘universal’ B0 shims, more information about which can be found in this ISMRM abstract: /universal_shim_-_ismrm_2025.pdf
Those universal shims have the advantages of being robust (no need for acquisitions or computations) and fast (instant), and can be used as either a stand-alone or as an improved initial estimate (see next section). As a stand-alone, they should be totally fine for spin-echo based structural scans and most gradient-echo based acquisitions, and may even provide very reasonable results for EPIs.
The universal shim can be loaded by selecting the ‘Thijs - Load Universal B0 shim (brain)’ (assuming you're scanning a brain) option in the PRIDE drop-down menu. The following pop-up will appear for confirmation, after which you're good to go.
¶ Universal B0 shim + subject-specific fine-tuning
The universal shim described in the section above yields much improved homogeneity compared to the unshimmed case. However, no two people are the same, so subject-specific shimming can generally still yield better results. But in the conventional subject-specific shimming approach, the substantial inhomogeneity can result in unresolved phase wraps, which can have a large negative effect on the accuracy of the calculated shim. Therefore, it's more reliable to start off with the universal shim as an improved first estimate, and calculate a subject-specific shim as a fine-tuning step to this universal shim. The figure below shows an example of the improved shim performance when using this approach (b) instead of using the standard approach in case of phase wrap (b).
For this, you can use the “B0 UniversalShim” sequence in Hospital→General→default (which can be seen in the figure in the ‘Getting started’ section above). This includes a B0 scan which uses the (hard-coded) univeral shim, followed by two Compass jobs:
To get a good shim, it is essential that both PRIDE jobs in (a) finish successfully. The first job uses Compass to calculate the modifications to the universal shim which will yield the best subject-specific results, after which the second job adds these modifications to the universal shim coefficients. The final results are written to the location where the scanner automatically reads them, and are printed into the terminal window shown in (b). Note that as the Compass job in the first step only calculates the differences relative to the universal shim, the coefficients shown in the Compass windows do not correspond to the final shim coefficients which you'll find in the scanner interface.
This tool is only built for the brain-based universal B0 shim, not for other anatomies.
¶ Troubleshooting
¶ ProcessingService crashed
ProcessingService errors appear relatively frequently. However, in some cases, PRIDE jobs do still work. If that's not the case, there are a few things you can try:
- On the ‘ProcessingService crashed’ pop-up, click ‘Retry’ to see whether the problem resolves;
- Open a terminal window and type ‘permproc stop ProcessingService’ followed by ‘permproc start ProcessingService’;
- Stop active jobs (by right-clicking them on the protocol list, and/or through the Administration window);
- If these don't work, close the scanner interface ‘Exit system’ and restart it. In some cases it’s best to make a new(!) patient, not to open the previous one. Then try again in this new patient.
- In all cases, if you are unsure or need assistence, feel free to call support.
If the PRIDE interface doesn't work but the scanner itself and Compass both work fine, for e.g. B0 shimming and MP2RAGE you can also circumvent the PRIDE interface:
- Acquire your data (e.g. B0 unshimmed)
- Export the XML/REC files through the Administration window (follow the steps in https://wiki.spinozacentre.nl/en/data_export, but select XML/REC instead of PAR/REC)
- Open the Compass GUI
- Select ‘Tools’ → ‘B0 shimming’
- Select the correct data path for the data you've exported in Step 2, and adjust the masking parameter if needed.
- Click ‘Run Tool’, and wait until the job finishes (as described above).
- In this case, take a bit of extra care to check that shimming has gone well: was the correct subject's data used, and did masking work properly?
¶ PRIDE job started, but takes forever
There is a weird bug in PRIDE where a job sometimes stalls without crashing. This happens more often if you've opened the corresponding terminal window and clicked anywhere on it, but it also (infrequently) happens at random times. Although this is a bit annoying, the fix is fortunately very easy: click anywhere on the stalled terminal window, and hit ‘Enter’ and/or the space bar (I generally just do both to be sure…). This should instantly get things running again.
¶ COMPASS opened, but nothing happened
If Compass opened after you started a Compass PRIDE job, it probably means that you forgot to open it at the start of your session (or multiple parallel instances might have been running). In this case, the most reliable quick fix is to close all Compass windows, open Compass.exe again, and wait until it has fully loaded ("Idle" appears at the top of the Compass GUI). Then, you can re-run the Compass job (i.e. hold-drag-and-drop it onto the sequence in the protocol list).
¶ If everything fails…
If for whatever reason you really can't get a B0 shim to be calculated, try to use the Universal B0 shim (described above) as an alternative option. At least for brain MRI, this generally works much, much better no shim at all, and may well give you a useful dataset instead of having to cancel your session.
However, if the entire PRIDE interface is dead, it might also be that the ‘Load universal B0 shim’ PRIDE job won't work. In that case, you can also navigate to the G:/Patch/Pride/ folder and run ‘Thijs_UniversalB0_brain.dat', which should still work even if PRIDE is down. As a check, go to the protocol editor and check whether the following shim values are correctly loaded (under the ‘Contrast’ tab):
X = -0.0313
Y = -0.0024
Z = 0.1143
Z2 = 1.3072
ZX = -0.1894
ZY = 0.0389
X2-Y2 = -0.1815
2XY = 0.0324