Long Delay with Superlab 5


I’m having difficulty with this experiment. When I try to run it, I get a long lag, both when trying to name/save the data at the beginning of the experiment, and between image presentations. The presentation of the same images work exactly as they are supposed to, there’s no problem with that portion of the experiment, nor is there a problem/lag when participants hit the space bar to denote that they have detected a change. The lag occurs when they hit enter after explaining what the change was. The wait for the next picture presentation to occur is between 13 and 20 seconds every single time. Any ideas? Thanks!

Lindsey Change Detection.zip (65.1 MB)

Lindsey Change Detection.sl5 (357 KB)

The problem is in the way the experiment is designed. The main trial, “Pre/Post Change Scenes”, is linked to about 400 events, half of which use lists and each list has 49 items, so SuperLab is trying to load 200 times 49, or about 9800 picture files.

How would you like trials to be presented? Perhaps I can offer an alternate approach.

Thanks for your reply. That would certainly explain the lag. For this experiment, I need to present a picture from the prechange scenes stimulus list for 240 ms, followed by an ISI of 80 ms, then repeat the prechange scene for 240 ms followed by another ISI of 80 ms. After that, I need to present the corresponding postchange scene twice for 240 ms, separated each time by an 80 ms ISI. So prechange-ISI-prechange-ISI-postchange-ISI-postchange-ISI, and this pattern needs to repeat for 75 seconds or until the participant indicates that they are aware of the change by hitting the space bar and typing in their response. Hopefully there’s an easier way to do this so that the computer doesn’t have to load 9800 stimuli every time!

Thanks again for your help,

Hello Lindsey,

The experiment is indeed feasible, with a small caveat. The attached experiment will present the sequence “prechange-ISI-prechange-ISI-postchange-ISI-postchange-ISI” and repeat it as long as the participant has not pressed a key within 75 seconds, but there is a perceptible delay in between sequences.

The reason for the delay is this: SuperLab does its best to present events within a trial as accurately as possible. This means that any “housekeeping” chores that need to be done, e.g. loading pictures, allocating memory, etc., are done before or after a trial, but never during a trial. In your experiment, this sequence is implemented as a trial that gets repeated, hence the inter-trial delay. If you can live with this small delay, then your experiment is ready to go!

Other notes about the experiment:

  • Randomization is done in the trial’s [B]List Access[/B] tab, not at the block level.
  • I changed the ISI events to become events of Type text, with an 80msec time limit. The effect is the same, but this change SuperLab to detect a participant response during that 80msec pause.
  • You need to use SuperLab 5.0.3, which is the most recent version. If you have an earlier version, you can download the latest SuperLab 5.x update. The experiment uses a parameter of type Timer; this feature was added recently in version 5.0.3.
  • While working on your experiment, I discovered a bug that happens when SuperLab loads an experiment and is unable to locate stimuli files, and then prompts you to locate the files. The bug is caused by the fact that your stimuli files in the two folders have identical names. I have renamed the files in the "Postchange Scenes" folder to circumvent this bug.

Lindsey Change Detection rev B.sl5 (37.2 KB)

Postchange Scenes.zip (32.9 MB)

The bug that you run into that happens when SuperLab loads an experiment and is unable to locate stimuli files has been fixed in the recently released version 5.0.5.