crisglav / discover-eeg Goto Github PK

Hello :)

I'm thinking of adapting your 'compute_peakfrequency' function to additionally output localmax and cog PAF values for each channel to complement the global PAF you have already.

If this is something you already have let me know, otherwise I can share with you once I have it working and you can decide whether to have it in your software.

Best,
Sammy

Thank you for creating this EEG-BIDS pipeline. This looks promising!

We are planning to apply the pipeline to the Healthy Brain Network dataset from the Child Mind Institute. This dataset provides individual MRI T1 images, which were processed using Freesurfer.

Is there an instruction somewhere for us to use DISCOVER EEG with individual MRIs?

I saw these two parameters:
"HeadModelPath" : "standard_bem.mat",
"SurfaceModelPath" : "surface_white_both.mat"

How do we create these .mat files from Freesurfer/Nifiti files?

Also will the pipeline use electrode coordinates embedded in the BIDS files? See these coordinate files here
https://openneuro.org/datasets/ds004186/versions/2.0.0

Hi all,

I'm encountering an issue with my .bids data in discover-eeg, but always showed these error.
_Dot indexing is not supported for variables of this type.

Error in plot_badchannels (line 36)
chanlabels = {EEG(1).urchanlocs.labels};

Error in main_pipeline (line 207)
plot_badchannels(params,ALLEEG);_

What could be causing this issue? Thanks for your help!

Hello, thanks for making this pipeline!

I was hoping to change the epoch length and run your script. But it seems that you must specify a params.FreqRes that is higher (i.e. lower number) than the default for the params.EpochLength you set (i.e. 1/EpochLength).

For example, if I set params.EpochLength to 5, and params.FreqRes to 0.2 or [ ], (i.e. the default that is calculated if left empty), then calling the function define_params.m does not work. I believe this is because the logical statements within define_params.m to check the params.Pad parameter do not work unless params.Pad has been specified manually or becomes set to 1/params.FreqRes at line 248 when the FreqRes has been assessed to be smaller than 1/EpochLength. The latter is the case for EpochLength of 2 and FreqRes of 0.1 as was set in your params_example.json.

I have set up a workaround for my self by commenting out the check for params.Pad and having no padding specified, but this might be something for you to look into. And of course, if you can't replicate the problem, let me know, as there might be other issues happening on my end if so :)

A small extra thing in relation to this, within your compute_power.m function line 12, you have cfg.pad = params.Padding, which I believe should be cfg.pad = params.Pad to match the params_example.json.

Best wishes,
Sammy