Loading CND format dataset with load.cnd about eelbrain HOT 4 CLOSED

Hi @xinyiguan, I contacted Giovanni Di Liberto about what the desired behavior should be if origTrialPosition is missing.

from eelbrain.

I haven't heard back yet, can you maybe tell from the data what should happen? Can you do the analysis without origTrialPosition, so that it would be okay to just drop it?

from eelbrain.

I resolved this by adding the origTrialPosition values myself. Thanks for stepping in!

from eelbrain.

Great! @diliberg commented

that variable should not be empty. That is an issue with the dataset. I did that quickly for the workshop, I am afraid, and I forgot to fix it as it was not important.

In brief. Let me give you two examples:

1. audio-book listening experiment. The trial order was the same for all participants (as the order was important. So, origTrialPosition should be 1:nTrials.
2. Bach melody listening. The trial order was randomised -> different for different participants. So, when building the CND file, the approach I used was to sort everything according to a particular order (audio 1, audio 2, etc), so that trial 1 corresponded to audio 1 for all participants. origTrialPositionis a variable that allows us to remember in which order that trial was actually played for a given subject. I'll fix that when I have some time, but that should not be a problem (exactly because the order was randomised).

So from the next Eelbrain release there will just be a warning.

from eelbrain.