Random sampling and repeat presentation of stimuli

I’m setting up a face recognition expt. which has two main phases; a learning phase and a test phase. In the learning phase I need to select, at random, a total of 64 faces (8 faces each from 8 different sets -stimulus lists- of 64 faces). The faces are presented in turn after a fixation cue. The participant studies the faces. No response is necessary. In the test phase I want to present these same ‘seen’ faces along with another randomly selected set of 64 ‘unseen’ faces from the 8 stimulus lists. The participant then has to respond according to whether or not he recognises the faces from the learning phase.

  1. Is it possible, within SuperLab, to select a random sample of ‘seen’ faces to be shown in the learning phase and then have that same sample repeated (hopefully in a new random order) in the test phase?
  2. Select a new random sample of ‘unseen’ faces, EXCLUDING the faces that were selected for the learning phase?
  3. If the above is possible can I present the ‘seen’ and ‘unseen’ faces within the same block in a random order?
    I shall look forward to seeing your response!:slight_smile:

Not that I know of.

The way I typically do it is list the 640 or however many pictures in excel, use =rand() in a neighboring column, then sort it all by that column. Then group those randomly sorted 640 pictures into 10 groups of 64 (i.e., 1-64, 65-128, etc). Then make 10 different superlab scripts, one for each group of 64 (which were randomly selected).

The easiest way is to put all of the pictures (all 640 of them) into superlab, all ready to go as though you were going to run the study with that many images. Then in each script select only the 64 events that you want presented in that script (as determined via your random sort in excel). Then within each script make superlab present the 64 learning events in random order, and the 64 test events in random order.

Importing from Excel

Hi ricehead,
Thanks for your reply. I can sort the face filenames (jpg files) in Excel but have so far failed to import them into the Superlab stimulus lists because the events are individual picture files not text. How do I import the randomly sorted jpg files listed within Excel as picture files within Superlab?

Yes to all of the original questions, but it’s not obvious. I’ve attached an experiment to demonstrate this at a very simple level so far as I’ve understood the request.

I’m going to refrain from laying out step-by-step directions, as it was fairly detailed, but I will point out the important places to look within the attached experiment:

“When to Run” tab in the Block Editor
“Macros” tab in the Block Editor
“Randomize” tab in Block Editor

Stimulus lists and their code values

“When to Present” tab in the Trial Editor
“Code Values” tab in the Trial Editor

Experiment Menu->Codes. Note that most of these codes are on trials and not events.

“Input” tab in Event Editor
“Correct Responses” tab in Event Editor
“Feedback” tab in Event Editor on “Seen Stimuli”

Ultimately, the catch is that the original set of trials presented must be exactly the same trials used later. We use feedback and macros to set code values so that we can determine whether trials should be presented and what the correct response should be. The code values are used as “variables” in the mathematical or programming sense in order to affect the behavior of the experiment.

Note that because we’re using the “When to Present” tab in the Trial editor, you must be using a version of SuperLab that has this feature (4.0.6 or later, I believe) before solving your problem is even possible.

SubSamples.sl4 (10.6 KB)

Thanks for this. Now I have version 4.0.7 I think I can do what you are suggesting to make my experiment work. It’s a whole new level of complexity for me but it’s great to see what the new version of SL4 can handle.