[Eeglablist] DIPFIT wiki tutorial still broken?

[Curtis Bingham]

Arno, 

Thank you for updating the first part of the tutorial. Unfortunately, I'm still finding a large number of differences in the tutorial vs. actual. With these differences I cannot tell whether or not the dataset, processes, or software are incorrect.

 

I’m using Matlab R2019b, EEGLAB2019_1, dipfit v3.3, and the data set downloaded as per the instructions in the wiki at https://sccn.ucsd.edu/wiki/A08:_DIPFIT

 

I’ve carefully followed the instructions up to the section labeled, “Avoiding co-registration by choosing appropriate channel locations”.

 

However, the channel locations don't match up in the Edit Channel Information Info window. 

 

The tutorial image is here: https://sccn.ucsd.edu/wiki/File:Dipfit_pop_chanedit.jpg

But the actual dialog image is here: https://imgur.com/BC39Rmd

 

By default the channel displayed in the actual is OZ, not FPz like in the tutorial. 

Scrolling through the locations to reach FPz, the coordinates are very different than the tutorial image: https://imgur.com/TRfG7NI

All of the fields are very different, not the least of which is the number of channels: 32 in the tutorial and 69 in the actual dataset. 

After following the instructions to: “Press the Look up locs to look up your channel locations (by matching the channel labels) in the template channel location file.”

the Edit channel info window looks like this: https://imgur.com/kJztzve and still looks nothing like the tutorial image.

 

Continuing with the tutorial to the section titled, “Initial fitting - Scanning on a coarse-grained grid”,  the Coarse Fit (grid scan) window is different in that it shows 71 components vs. 32 in the tutorial:

Tutorial Image: https://sccn.ucsd.edu/wiki/File:Pop_dipfit_batch.png

Actual dialog image: https://imgur.com/s3u7bCg

 

The computation of the electrode projections shows a number of errors in the command window. It is unknown if this is expected behavior. 

https://imgur.com/bsXYVLH

 

The biggest and most disturbing problem shows up in plotting the dipoles.

 

Tutorial image: https://sccn.ucsd.edu/wiki/File:Dipplot_gridsearch.gif

Actual dialog image: https://imgur.com/jqZbQOa

 

I get 48 dipoles in my plot instead of 47 in the tutorial. My plot shows very different locations than the tutorial image shows.

 

Continuing further to the section labeled, “Interactive fine-grained fitting”

The Component to fit edit box shows differences:

Tutorial image: https://sccn.ucsd.edu/wiki/File:Pop_dipfit_nonlinear.gif

Actual dialog image: https://imgur.com/bfqA0XH

The residual variance is very different, and the positions, moments for dipole #1 are very different. 

 

When plotting the map, I get a dipole on my map where yours doesn’t have one: 

Tutorial image: https://sccn.ucsd.edu/wiki/File:A24comp3.gif

Actual dialog image: https://imgur.com/HDenGI6

And mine has a scale on the right where the tutorial does not.

 

Plotting the two dipoles, there are additional, significant differences.

Tutorial image: https://sccn.ucsd.edu/wiki/File:Diplot_image2.gif

Actual dialog image: https://imgur.com/V0rmFw8

 

When I try the Automated dipole fitting, I get an error:

https://imgur.com/CszQs1F

 

Can anyone tell me what is going on? Is the tutorial out of date? Can you reproduce the exact image with the tutorial instructions?

 

Thanks,

Curtis