Problem visualizing EOG and EMG of .edf file (normal eeg tracks view)

mcataldi · May 24, 2018, 8:53pm

Hi everyone,

I'm tryng to analyze an .EDF file registered from an EBNeuro machine. 22 channels (16EEG, 2EOG, 2EMG, 1 ECG, 1 edf annotations)
I can normally view EEG tracks, but nothing for both EOG and EMG and ECG.

I edited manually channels

Below an other screenshot:

I noticed that in the right amplitude band of EOG/EMG is shown "no units", maybe for someone expert could be useful.

I apologize, but I'm new in EEG source analysis and in Brainstorm Software; I checked in tutorials and in past topic, but i couldn't find solutions.

Thank to everyone will spend time to help me.

Greetings to all,

Matteo

Francois · May 25, 2018, 9:16am

Hi Matteo,

Are you sure you have signals correctly recorded in these channels?
Does it look any better if you display only one channel type at once (eg. only EOG).
“No units” means that the units could not be identified automatically (in uV for instance), typically either because there is no signal at all or because the values are too high to be physiological signals (> 0.1V for instance).

Another source of problems with EDF files are mixed sampling rates. I don’t remember if Brainstorm handles this correctly. Is there any chance your EMG and EOG were recorded at different sampling rates?
You can check this by looking in the header of the .edf files, saved in the links in the brainstorm database. Right-click on the “Link to raw file” > File > View contents.

If you really don’t understand what is going on or your suspect Brainstorm to behave incorrectly, please send us an example file so we can have a look at it. Upload it somewhere (google drive, dropbox, etc.) and post the link here.

Cheers,
Francois

mcataldi · May 27, 2018, 3:23pm

Hi Francois,

first of all thanks for your time.

I think signals (EOG, EMG, EKG) are correctly recorded (can normally view on Galileo suite and other sleep staging software).

Actually EDF file has different sampling rate (EEG and EMG 256 Hz, EOG and ECG 128 Hz), I had to set values down due to limited storage volume of our old portable recording hardware.

Next week I’ll try to analyze a file with same sampling rate, otherwise I’ll try to export tracks in another format (in hypothesis were a specific EDF problem).

If both tests will fail, I’m gonna update you and send an example file.

Thank you so much,

Matteo

Francois · May 28, 2018, 6:04pm

I checked the code, and channels with lower sampling rates are indeed ignored when reading from EDF files in Brainstorm. This is probably why you don’t see anything in the EMG and EOG windows.

github.com

brainstorm-tools/brainstorm3/blob/master/toolbox/io/in_fopen_edf.m#L146


    else
        error(['The number of samples is not specified for channel "' hdr.signal(i).label '".']);
    end
end
hdr.signal(i).sfreq = hdr.signal(i).nsamples ./ hdr.reclen;
end
% Find annotations channel
iAnnotChans = find(strcmpi({hdr.signal.label}, 'EDF Annotations'));  % Mutliple "EDF Annotation" channels allowed in EDF+
iStatusChan = find(strcmpi({hdr.signal.label}, 'Status'), 1);        % Only one "Status" channel allowed in BDF
iOtherChan = setdiff(1:hdr.nsignal, [iAnnotChans iStatusChan]);
% Remove channels with lower sampling rates
iIgnoreChan = find([hdr.signal(iOtherChan).sfreq] < max([hdr.signal(iOtherChan).sfreq]));    % Ignore all the channels with lower sampling rate
if ~isempty(iIgnoreChan)
iOtherChan = setdiff(iOtherChan, iIgnoreChan);
end
% Get all the other channels
if isempty(iOtherChan)
error('This file does not contain any data channel.');
end
% Read events preferencially from the EDF Annotations track
if ~isempty(iAnnotChans)

Brainstorm can’t handle signals with mixed sampling rates. It would be possible to reinterpolate the signals with the lower sampling rates using a higher rate, but it would require some consequent rewriting and testing of the current reading functions.
If you are interested in working on this, you are more than welcome than welcome to contribute to the project!