[Eeglablist] Problems with precomputing channel measures in EEGLAB 2021.0

Malte Anders malteanders at gmail.com
Thu Apr 15 03:35:35 PDT 2021

Dear EEGLAB List,

I have a couple of datasets that I analyzed using the STUDY function of
EEGLAB 2020_0 without any problems at all. I load the datasets into EEGLAB
using my script and then precompute ERP, ERSP etc. using the STUDY menu.
It's ~50 subjects and a couple of conditions and groups. This all works
like a charm in 2020_0.

I just recently "updated" to EEGLAB 2021.0. All works well until I hit the
"Ok" button in the Precompute Channel Measures. After that, exactly nothing
happens. In the MATLAB command window, nothing is shown like it was in
2020.0 (usually, the process of "Reading float file [xxx]... Saving ERP
file [xxx]" would begin, but this is not the case in 2021.0). When I hit
Control + c to stop MATLAB, I get the following error over and over again
(if I don't hit Control + c, the output would be a million lines long):

In eeg_getdatact (line 150)
    EEG = pop_loadset('filename', EEG.filename, 'filepath', EEG.filepath,
'verbose', 'off');

In eeg_checkset (line 683)
            EEG.data = eeg_getdatact(EEG);

In pop_loadset (line 221)
        EEG = eeg_checkset(EEG, 'loaddata');

Please note that after this, ALL datasets have gone bad and cannot be
loaded into EEGLAB even as a single dataset. Basically, EEGLAB damages the
file so that File -> Load Existing Dataset gets stuck at:
pop_loadset(): loading file [xxx]
and the EEGLAB GUI does not update.

As I keep backups, this is not a big deal, but if I go back to 2020_0,
everything works again like a charm.

I think there might be a bug in 2021.0 when it comes to dataset loading in
the study function. Maybe it has something to do with the fact that now 1
instead of 2 files are saved by default? Keep in mind that the files I try
to load into the study function consists of two files .set and .fdt, I
don't know if this has something to do with this.

For now I will go back to 2020_0. I'd like to use the new EEGLAB - maybe
somebody can help?


More information about the eeglablist mailing list