Although ICLabel automated rejection allows to chose a threshold interval for each category probability, the implementation on Automagic currently permits to only insert a one-value threshold, above which the component is either included or rejected. This might be disadvantageous if one wants to set a x < thr together with x > thr exclusion/inclusion criteria.
I tried myself to overcome this by applying a few changes in the scripts, that are:
\gui\settingsGUI.m : from line 530 I replaced “str2double” with “str2num”, so that arrays are also supported as string input.
\preprocessing\performICLabel.m : from line 122 I added a further test on the second element of the “*Ther” variables (e.g. see attached)
I tested the pipeline with the changes, and it seems to work smoothly, although there is a problem with the GUI, as the text boxes are not large enough to display a two elements array. I reckon that this can be easily fixed though.
I was wondering whether you may kindly provide me with a feedback on my fixes. Specifically, has this particular feature already been tested during the development? If that is the case, would you by any chance have any contra-indication against my fix (i.e. possible known bugs produced by such changes)? Is there any reason why you deemed including one only threshold value more efficient?
Dear developers,
Although ICLabel automated rejection allows to chose a threshold interval for each category probability, the implementation on Automagic currently permits to only insert a one-value threshold, above which the component is either included or rejected. This might be disadvantageous if one wants to set a x < thr together with x > thr exclusion/inclusion criteria.
I tried myself to overcome this by applying a few changes in the scripts, that are:
str2double” with “str2num”, so that arrays are also supported as string input.I tested the pipeline with the changes, and it seems to work smoothly, although there is a problem with the GUI, as the text boxes are not large enough to display a two elements array. I reckon that this can be easily fixed though.
I was wondering whether you may kindly provide me with a feedback on my fixes. Specifically, has this particular feature already been tested during the development? If that is the case, would you by any chance have any contra-indication against my fix (i.e. possible known bugs produced by such changes)? Is there any reason why you deemed including one only threshold value more efficient?
Many thanks in advance!
Ramtin