Inconsistency in 'Fitting' Results Between 2012 and 2025 – Seeking Guidance

[BoyuanSong]

Dear Denis Brunet,

I hope this message finds you well.

I am a PhD student at UCL currently working on a project involving topography-based analysis. As part of my research, I utilize averaged epileptiform spikes to generate topographical IED maps, which are then correlated with EEGs recorded inside the scanner using the 'Fitting' procedure in Cartool. In this process, spikes are averaged, and the EEG voltage map corresponding to the peak of the Global Field Power (GFP) of the average spike is considered as the epileptic map(template for 'fitting').

While re-analyzing cases originally processed in 2012, I noticed that the 'Fitting' results differ when re-running the same cases in 2025. To illustrate the issue:

Figure 1: 'Fitting' result from 2012

Figure 2: Corresponding .vrb file from 2012

Figure 3: 'Fitting' result from 2025

Figure 4: Corresponding .vrb file from 2025

I would greatly appreciate any insights you might have regarding potential reasons for this discrepancy. Could this be related to changes in parameter settings, or might it be due to updates in Cartool over the years?

Thank you very much in advance for your time and support.

Kind regards,
Boyuan
Figure 1. 'Fitting' result from 2012

Figure 2. Corresponding .vrb file from 2012

Figure 3. 'Fitting' result from 2025

Figure 4. Corresponding .vrb file from 2025

[DenisBrunet]

Hi,
Thanks for reporting...
For sure code has evolved a lot in 13 years! Still, we should somehow get similar results, if not identical.
After a very quick glance I can see at least one parameter that is different, which is the Smoothing one. It is off in 2012 and on in 2025.
There might be more, though, like the Spatial Filtering which is not hinted in the 2012 snaphsot.
Also I am a bit puzzled by the different GFP values, as well as the weird correlation values, which also hint that this is not a regular EEG...

I can have a quick look with some data of mine with some similarly old version.
Best,
Denis

[BoyuanSong]

Hi Denis,

Thank you very much for your response. You're absolutely right—this EEG dataset was recorded simultaneously with fMRI, so it may contain a significant amount of artifacts.

I’ve plotted a comparison of GFP and correlation for data from 2012 and 2025. Please find the charts below:

Series 1 (blue) represents the 2012 results
Series 2 (orange) represents the 2025 results

Both the GFP and correlation values appear to follow a similar trend. However, I noticed that a few of the correlation values(in 2025) are exactly zero. I'm wondering if a correlation value of zero is expected or indicative of an issue in this context.

If you happen to have a chance to check one of your datasets for comparison, it would be greatly appreciated.

Thanks again for your support.

Best regards,
Boyuan

[BoyuanSong]

Hi Denis,

I hope you're doing well.

I just wanted to follow up and kindly ask if there have been any updates from your side regarding the test with your older data version. I completely understand that your schedule is busy, but I’d really appreciate any insights you might have been able to gather—particularly concerning the differences in GFP values and the unusual correlation results.

Thank you again for your time and support.

Best regards,
Boyuan