Different stimulus parameters on every repetition

[celestialteapot]

Hello,

I created an experiment where the stimulus parameters are computed by a function (multiple random samplings involved; three stimuli without interaction, then a test) and I would like to run several of this trial sequence consecutively. Crucially, each repetition should have different parameter values (ie. a new output from the function), but by using the 'repetitions' parameter I get the exact same value set for every rep. I passed the parameter-value-computing function to 'timeline-variables' as 'timeline-variables: my_func()'. Is there a solution for this problem?

As an alternative resolution I tried using the 'call-function' plugin and passing its 'value' to the stimulus trial, but this doesn't work either, since it gives undefined for the 'call-function' trial data. If I understand correctly the timeline is compiled before running (or something like that), so the output of the 'call-function' trial does not exist yet when the stimulus trial would already want to have it?

I guess this topic is similar to (the same as?) the dynamic parameters/timeline variables problem.

Edit [clarification]: my timeline would look like [fixation, regenerate_stimuli, stimulus (x3), fixation, test] repeated multiple times.

Thanks for this awesome library,
Balint

[becky-gilbert]

Hi @celestialteapot, sorry for the delay. It's hard to understand exactly what you're trying to do and what problems you're having without seeing your code. Would you be able to share it please (even if it doesn't work)?

Here's some general information, in case it's helpful.

You're right that with a timeline_variables procedure and the repetitions parameter, you will get the exact same parameter values from your timeline_variables array across repetitions. If you want your parameter values to be different on each trial, and you don't know in advance what the values should be, you can use a function as the parameter value. Just make sure that you either use an anonymous function or a function name (without the parentheses at the end) so that the function is called right before the trial starts (rather than when the page loads).

Here's an example of an anonymous function for the stimulus parameter:

var feedback = {
  type: 'html-keyboard-response',
  stimulus: function(){
    var last_trial_correct = jsPsych.data.get().last(1).values()[0].correct;
    if (last_trial_correct) {
      return "<p>Correct!</p>";
    } else {
      return "<p>Wrong.</p>"
    }
  }
}

And the same thing but with a named function:

function generate_stimulus() {
  var last_trial_correct = jsPsych.data.get().last(1).values()[0].correct;
  if (last_trial_correct) {
    return "<p>Correct!</p>";
  } else {
    return "<p>Wrong.</p>"
  }
}

var feedback = {
  type: 'html-keyboard-response',
  stimulus: generate_stimulus
}

And if you want to include information in these functions that comes from the trial information stored in timeline_variables, you can do that by using the jsPsych.timelineVariable function with the second true argument inside of your parameter function. The code below is taken from the second example in the timeline_variables docs:

{
  type: 'html-keyboard-response',
  stimulus: function(){
    var html="<img src='"+jsPsych.timelineVariable('face', true)+"'>";
    html += "<p>"+jsPsych.timelineVariable('name', true)+"</p>";
    return html;
  },          
  choices: jsPsych.NO_KEYS,
  trial_duration: 2500
}

Sorry if none of this helps with your question... please feel free to post some code/examples and we'll do our best to offer more specific suggestions!

[jodeleeuw]

Hi @idoba2030 ,

Thanks for creating a separate discussion post, #1313, to track your question. We prefer to have one main question per thread to help folks find solutions through search.