Reaction time and feedback

Hello,

We wanted to prepare a Classic Conditioning experiment using a Reaction Time as US task, however, I do not know how to program it in SuperLab 4.5.

In the experiment, a CS + of 5 second duration is presented (a simple geometric figure) and immediately after this stimulus a 0,4 s noise is presented. The participants have to respond by pressing the space bar as soon as they hear it.

In addition, we want to provide positive feedback, so that in cases where the participant responds quickly than the previous trial, a message of reinforcement appears on the screen (e.g you respond faster than the last trial), and no message when the response time is longer than the previous trial.

The problem is that I don´t know how to schedule a Reaction time task or feedback, if anyone can help me it would be great.

Thank you very much.

Go to the Feedback tab, and select the reaction time is too quick as a criterion. You can then present an event as a consequence of response time being below a certain threshold.

However, to satisfy your requirements, your threshold would need to be variable. Unfortunately, this cannot be done with SuperLab 4.

Please note that SuperLab 5 consists of timer parameters that would allow this behavior.

Reaction time + Variable Feedback

Hello again,

As I previously said, I want to rule out a Classic Conditioning experiment using a Reaction Time as US task. I have now the Superlab 5 version, but I still don´t know how to schedule the Reaction time+Feedback procedure.

In this experiment, a 5 s. CS + is used (a simple geometric figure) and immediately after this stimulus a 0,4 s. noise is presented. The participants have to respond by pressing the space bar as soon as they hear it. We want to provide feedback only in that cases where the participants respond quicklier than the previous trial. We want to show them a message of reinforcement on the computer screen (e.g you respond faster than the last trial), and no message when the response time is longer than the previous trial.

Thank you very much.

I’ve attached an experiment that demonstrates this behavior. Please read through the notes.

The experiment revolves around 3 parameters: two timers and one text parameter that determines which timers have measured the current or previous trial.

Let me know if you have any questions.

rtFeedback.sl5 (6.82 KB)

Hello again,

We have had success with the configuration that sent us, however, our experimental design requires a second CS- (blue triangle) stimulus and the already mentioned CS + (red square), also the interval between stimuli must be variable. When the experiment is started without CS- the timers work well, the feedback is displayed properly (with the criteria dictated above). However, if we add the CS- the results are not consistent, since feedback is not shown randomly and does not meet the pattern we want.

We do not know what the error is or how to configure it for the timers to ignore the CS-.

If this is not possible, another option could be the one carried out by Lipp and Vaitl (1990), which consists of comparing the mean RT of the last responses with the current RT. This is reflected in the papper on page 78: “For the RT-new group only, the subject’s RT Was shown on the television screen, and enhanced Feedback was given. Whenever the current RT was Faster than the mean RT during the previous five Trials, a difference score was displayed together with Positive feedback (RT <mean, “good”, 20 ms <D <40 ms, “very good”; D> 40 ms, “excellent”).
Subjects were not informed about the scaling of The various feedbacks provided. Feedback information Lasted for 3 seconds on the screen. The Subject’s response was a little button placed Underneath the left index finger as soon as the Imperative stimulus was sounded.
” (see Lipp, O. V., & Vaitl, D. (1990). Reaction time task as unconditional stimulus. The Pavlovian journal of biological science, 25(2), 77-83.)

Thank you again for your help.

Have you tried presenting the CS- in its own trial in the same block? See the experiment I’ve attached. Does this suit your requirements?

rtFeedback2.sl5 (7.57 KB)

The problem still persists, since in my procedure I need to introduce a variable interval-between-trials (ITI) and a randomized presentation of CSs. The RT that I want to measure is after finishing the CS + (specifically after the presentation of a tone presented immediately after the CS+) and give feedback if it is smaller than the RT to the previous trial. The error that I find is that the RT that is measured does not correspond to reality, since sometimes that criterion is not met (sometimes it gives feedback by responding more quickly than the previous trial but at other times not, and vice versa ).

I think that may be due to both CS- and ITI as follows: in an example of CS + sequence followed by another CS +, if we respond slow at first (e.g 5 sec) and very quickly after the second CS + (1 sec. ), the feedback appears. So far all correct, however, if we continue the sequence with a CS- and then a CS +, although after this CS + we respond faster than in the previous CS + (here 0.7 s instead of 1 s.), we do not receive feedback.

Thus, we believe that the RT can be measured from one bar pulse to the next one, instead of measuring the RT after the tone presented immediately after the CS +. In this way it would be understood why when a CS- is presented, that is to say, the interval between the CS + and therefore of pulsations is extended, no feedback is sent.

I attached an example of the procedure in superlab to see how it is produced.

PS: the response has to be immediately after the tone, and it can be within the interval of the tone, e.g if the tone lasts 0.4 sec. and if the participant responds to 0.3 sec, the answer would be worth it.

Thank you again for your help.

ExperimentFeedback.sl5 (8.78 KB)

I have divided the block into three trials: ITI, CS+, and CS-. I have also added a counter parameter (0 min, 1 max). Every time the ITI is presented, its trial rules randomize the counter and start the timer only when necessary. Depending on the randomized value and the When to Present options of the trials, only one of the CS+ or CS- trials will be (randomly) presented.

This should allow for: the timers only measuring CS+ trial; random presentation of stimuli; an ITI before each CS.

Take a look at the experiment attached.

ExperimentFeedback.sl5 (9.56 KB)

We have tried the experiment and it works very well, but we are missing a couple of details that we do not know how to include in the experiment. It would be possible to establish a “rule” in the randomization to establish that no more than 2 identical trial appear consecutive? That is, no more than 2 CS + or 2 CS- can be displayed continuously. We would also like to make sure that each CS is only presented 10 times along the experiment (10 CS+ and 10 CS-, 20 in total)

In addition, we wanted to know if putting the ITI as variable (with trial levels of 1000ms, 2000ms and 3000ms) instead of fixed, the ITI rules would be affected, and therefore the timers.

Thank you again for your help.

I have added two counter parameters: Consecutive CS+ and CS- Count. Each time a CS is presented, it increments its own counter and resets the other. This allows us to use the When to Present options to make sure that they are not presented more than twice in a row.

Something similar is done with the new Total CS+ and CS- Count parameters to prevent a CS from presenting more than 10 times.

Varying the ITI time limit will not affect the timing process for the CS+ trial. I’ve created the trial variable for you.

rtFeedback4.sl5 (11.2 KB)

We have tried running the file and there are still some things we did not get them to work properly. In the first place, we want to establish exactly 10 CS + and 10 CS-, under the conditions previously mentioned. In the file, we have tried to change the maximum from <10 (previously established) to <11, trying to fix it, but the problem persisted (sometimes appear 7 CS + and 6 CS-, others 8 CS- and 6 CS +, 7 CS+ and 9 CS-, etc.). Our intention, as we have indicated, is that 10 CS + and 10 CS- must appear, randomly and with the restrictions previously indicated. Second, we have noticed that the ITIs may sometimes appear consecutively (eg lasting 3 s., others times lasting 6s, 8s, 9s, etc.). This should not happen because the variable was defined between 1 and 3 s.

In addition, we must note that we have selected “Present trials in this block = 20 times” in the tab “Looping” in “Block editor”, and we don´t know if it can be behind those problems.

Finally, we would also like to add a final message indicating the completion of the experiment.

Thank you again for your help.

You are correct. The stimuli do not reach 10 presentations because block looping was limited. I’ve changed it to infinite looping.

On that note, in order to stop the looping and present an end message, I’ve set criteria in the block’s When to Present tab: only present when total CS+ or CS- count is < 10.

I’ve also made other changes that will prevent your experiment from getting stuck and presenting ITIs consecutively. I’ve done this by adding rules to the ITI trial and removing a criterion from CS+ and CS- trials.

rtFeedback5.sl5 (13.1 KB)

The experiment is almost finished, it works perfectly whether it responds slower than the previous time (Feedback not shown) or responds faster than the previous time (Feedback message is displayed). However, we have tried to program it so that the time limit to which the participant can respond is 2 sec. after the start of the tone, and in case of no response during this interval, continue to ITI (something similar to the situation of “if it responds slower than the previous time”). The problem is that sometimes, when the participant don´t press the space bar, the Feedback message is displayed.

On the other hand, we also want that, when the participant gives a correct response (faster than the previous one and the Feedback is presented), the feedback message stays on the screen for 2 seconds and the participant doesn´t have to press the bar to continue.

If we could fix this problem, the experiment would be ready. To do this, we upload a copy of the experiment in superlab, and the CS and tone used, so you can test it under the same conditions.

Thank you again for your help.

ExperimentFeedback.sl5 (12.7 KB)

CS+ Red Square.PNG

CS- Blue Triangle.PNG

WN (0,4s).wav (34.5 KB)

See the experiment package attached.

I merged the tone and response into one event that lasts 2 seconds or until the space bar is pressed (correct response). If pressed, then the trial’s tag is set to correct; I’ve changed the rules so that the feedback only presents when this occurs.

I’ve also changed the time limit on the feedback event to last 2 seconds and end on its own.

rtFeedback6.zip (10.2 KB)

We have detected that the rule “no more than 2 CS+ or 2 CS- followed” sometimes doesn´t work in last trials, for example, when 10 CS+ have already been presented, but there are still 4, 5 or 6 CS- left, they appear consecutively, breaking the rule of “no more than 2 equal trials followed”. We would like to know if there is any way to satisfy this rule and keep the rest rules unmodified (10 trials of each CS, positive feedback, etc).

Thank you very much.

This occurs because of the random presentation of CS+ or CS- in concert with its limitations (max consecutive: 2, max total: 10). Are these not requirements of your design?

Yes, these are two requirements of our design, however, as I have previously indicated , sometimes the criterion of “no more than 2 CS + or 2 CS- followed” does not work well. For example, CS +, CS +, CS -, CS +, CS +, CS -, CS +, CS +, CS -, CS +, CS +, CS -, CS+, CS +, (until now, all the possible CS+ of the experiment have already been presented, ie 10, while only 4 CS-, so the sequence continues as follows) CS-, CS-, CS-, CS-, CS-, CS-. Thus, this shows how the rule “no more than 2CS+ or 2CS- consecutively” has been broken.

We would like to know if it is possible that it will not happen, maybe through a more balanced presentation of CSs over the course of the experiment. We suggest this because looking at the above example, if one kind of CS reachs the 10 presentations requirement very soon, the other can only reach this 10 presentarions requierement by breaking the rule “no more than 2 identical CS consecutively”). We think that a more balanced distribution of CSs throughout could help to keep both criteria unchanged, but we don’t know if this is the correct solution, and, if it is correct, we don’t know how to program it.

Thank you very much for your help.

This sort of balanced presentation will prevent randomization in portions of the experiment. There is only one way that a type of CS can catch up while not breaking the limitations that are set, and that is controlling it to present twice in a row while the other type is presented only once. I do not know your research goals or specifications so I don’t know if this is an appropriate solution to your design. Let me know and I will try to implement the solution into your experiment.

If the solution you propose works, try it out in our experiment. As I mentioned above, the rules of “exactly 10 CS + and 10 CS- presentations” and “no more than 2 CS + or 2 CS- consecutively” have to be mainteined. It is important not to break any of these rules as it is happening now (with more than 2 identical CSs presentations appearing consecutively in the last trials of the experiment). Indeed, when applied in this way, it would not be a randomization, but it would be a pseudo-randomization.

The rest of the conditions don´t have to change either.

Thank you very much for your effort and help.

If one CS is ahead of the other by a certain amount, then it is impossible to prevent your requirements from being broken. This point of no return seems to be when the remaining presentations of one CS is more than twice of the remaining presentations of the other; when Remaining CS1 > Remaining CS2 * 2.

I have created counters and rules that force-present CS1 (and CS2) just before this point; when Remaining CS1 = Remaining CS2 * 2

Please test the attached experiment.

rtFeedback7.zip (10.5 KB)