Millisecond Forums » Millisecond Forums » Inquisit 4 » adopted Columbia Card Task

adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 12:00:13 GMT

Hello,
I am totally new to Inquisit. I downloaded the CCT and made couple of modifications, such as reduced cards to 6. My questions are as following:
1. how to randomly choose cards from one of six cards by pressing "take card" button until subjects want to stop (by pressing the same button).
2.I can't control the trial duration (10 seconds) by setting /response=timeout(10000), /timeout or /trialduration in <trial pickcard>.

anyone can help me modify the syntax in the attachment, thanks.

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 12:00:13 GMT

Thank you very much for the kind words and I wish you much success with your research project!

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 11:24:12 GMT

Thanks Dave, you help me resolved all the (stupid) problems. I love you more than i can say.

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 10:51:04 GMT

> For example, if there were 3 loss cards in a round, the points feedback was correct when I pressed 1 or 2 times
> but 0 scoring appeared after I press 3times and stop pressing (do nothing and wait for timeout).

This is *precisely* what I pointed out in my previous reply. Again, look at:

/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.score-=values.score]

You've opened 3 "safe cards" and are now in the 4th instance of <trial pickcard>. Only 3 loss cards remain, i.e. values.cardtype at this point *is* 3 as in

/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.score-=values.score]

You do *not* respond in this trial. Hence, trial.pickcard.response *is* 0. Obviously, 0 is *not* equal to 8 as in

/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.score-=values.score]

It should now be obvious that the entire condition

/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.score-=values.score]

is met and the score is zeroed as instructed. In your particular case, if you *don't* want this to happen, change the logic to

/ ontrialend = [if(trial.pickcard.response=="7" && values.cardtype==3)
    values.score-=values.score]

Work through the rest of the logic accordingly.

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 10:35:26 GMT

Could you help me modify it, because I have to implement it a few hours latter. Thank you so much.

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 10:33:20 GMT

It is weired that it only happened when I turn over the last safe card. For example, if there were 3 loss cards in a round, the points feedback was correct when I pressed 1 or 2 times, but 0 scoring appeared after I press 3times and stop pressing (do nothing and wait for timeout).

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 10:04:51 GMT

Look at

<trial pickcard>
...
/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.score-=values.score]
/ ontrialend = [if(trial.pickcard.response=="7" && values.cardtype==2)
    values.score+=values.gainamount]
...
</trial>

That's where the score is manipulated depending on *response* and the type of scheduled *card* (loss or gain).

"!=" is a logical NOT. Thus,

/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.score-=values.score]

means "if the response is not a press on the key with scan code 8 AND the scheduled card is a loss card" subtract stuff. Now, which responses in your case are "not 8"? It's 7 and 0 (no response).

Similarly, there are other pieces of logic where you *don't* take into account the "no response" case:

/ ontrialend = [if(trial.pickcard.response!="8" && values.cardtype==3)
    values.openlosscards=1]
/ ontrialend = [if(trial.pickcard.response!="7" && values.cardtype==3)
    values.openlosscards=1]

Work through that logic, think through what exactly happens in each case and then modify it according to your needs, i.e., take the "no response" case into account.

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 09:38:16 GMT

I resulved part of problems, however, I still can't figure out how to deal with the problem of 0 scoring.

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 07:55:41 GMT

Re. the scoring: Work through the code that is responsible for that and double-check it. You need to decide what you want to do when no response occurs and adjust accordingly.

Re. the timeout: Again, just as with the BARTY, you ought to instruct the turnover trial to last for whatever remains of the 10s timeout. As with the BARTY, you do that via the /trialduration.

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 06:33:19 GMT

It appears to happening only when i press pick card button and then do nothing resulted in trial terminated automatically. Moreover, I alos found that when I press "stop" button the trial terminated immediately before 10s' timeout.

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 06:22:13 GMT

There is another problem. I found a discrepancy between two scores in CCT_hot. Specifically, sometime the score in pick card stage was consistent with that in feedback stage (e.g. 88 points in pick card stage and reveal the same score in latter feedback stage). However, sometimes these two scores were not identical (e.g. i got 100 points in pickcard stage but it revealed 0 points in latter stage).

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 05:57:41 GMT

Thanks :)

You just made my day!

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 05:53:27 GMT

Thanks Dave, I love you.

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 05:49:22 GMT

Just as with the BARTY, you ought to set the timeout variable to 10 seconds at the beginning of each round.

<trial getcondition>
/ ontrialbegin = [values.pickcardtimeout=10000; ]
...
</trial>

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 05:32:09 GMT

I revised the script as you mentioned in former post (please see the attachment). However, it only works on the first hot round, the rest of hot trials automatically and immediately terminated before i pressing the button.

RE: adopted Columbia Card Task

minin72704 — Fri, 14 Aug 2015 05:00:26 GMT

I add / branch = [if (trial.pickcard.sumlatency==trial.pickcard.timeout) trial.turnover] in <trial pickcard>, but it was not what I intended.

RE: adopted Columbia Card Task

Dave — Fri, 14 Aug 2015 04:59:16 GMT

You do it in the same way as with the BARTY, which we discussed in this thread.

RE: adopted Columbia Card Task

minin72704 — Thu, 13 Aug 2015 23:09:42 GMT

I finally found that the trial duration of CCT_hotwithin a round was more than 10s after I pressing more than once. Actually, i want one round terminate exactlyat 10s whatever how many times participants pressedwithin per round (e.g. In round 1, participant made two presses, first press was at 3s, secondat 6s, and i want it to terminate at 10s rather than 16s). Present syntax leads10s timeout to recalculate after each press within one round (please seeattchment). Could you help me resolve it? Thank you.

I finally found that the trial duration of CCT_hot within a round was more than 10s. Actually, i want one round terminate exactly at 10s whatever how many times participants press the button within that period (e.g. In round 1, participant made two presses, first press was at 3s, second at 6s, and i want it to terminate at 10s rather than 16s). Present syntax leads 10s timeout to recalculate after each press within one round (please see attchment). Could you help me to resolve it? Thank you. - See more at: http://www.millisecond.com/forums/Topic16800-7.aspx#sthash.zTWGB0iu.dpuf

RE: adopted Columbia Card Task

minin72704 — Thu, 13 Aug 2015 20:47:24 GMT

RE: adopted Columbia Card Task

Dave — Thu, 13 Aug 2015 03:59:44 GMT

> I know it works by setting trialduration to 9s

Then please use /trialduration. That's exactly the scenario it is intended for.

> but it will make participants fail to confirm their response while they press the corresponding keys

I'm not sure what that's supposed to mean. It literally makes no difference. With /trialduration specified, the trial goes into a posttrialpause for the remainder of the specified duration.

> How can it let it terminate at 9s whatever when participants make their responses within that trial?

Use /trialduration.

RE: adopted Columbia Card Task

minin72704 — Wed, 12 Aug 2015 21:10:23 GMT

Hi, Dave,
I really confused about the time control. I set a trial's timeout=9000 and / posttrialpause = trial.5losseva_self.timeout-trial.5losseva_self.elapsedtime (see the attachment). If the elapsedtime is 4s, the trial should pause 5s to the next trial. But it terminate 9s latter after my response, e.g. i response at 4s, it terminate at 13s. I know it works by setting trialduration to 9s, but it will make participants fail to confirm their response while they press the corresponding keys. How can it let it terminate at 9s whatever when participants make their responses within that trial?

RE: adopted Columbia Card Task

minin72704 — Tue, 11 Aug 2015 09:12:17 GMT

Thanks for your help.

RE: adopted Columbia Card Task

Dave — Tue, 11 Aug 2015 08:01:19 GMT

> It may be the last two questions (i hope so). First one, revised BARTY.iqx works well, however, it seems exceeding 2s than
> expect. Specifically, I expected the totoal time is 3mins 16s (1s fixation+10s instruction+15s*12 balloon+5s feedback).

Well, there's other stuff going on that you cannot control. Sometimes Inquisit will have to wait for the start of a display frame in order to start drawing stuff to the screen, etc. This can add up over the course of the entire task and there's nothing you can do about it.

As for the Cued Go/-No Go: Your math is wrong.

<trial cue> lasts 1300 ms plus a variable soa (see list.soa). <trial iti> lasts 700 ms as per values.iti. That alone gives you 2000ms + soa. If the end result is supposed to be approx. 4000ms for each (cue-target-iti) tuple of trials, the /trialduration of <trial target> should be set to 2000 ms - soa. I have no idea where you get the 3300 ms /posttrialpause figure from. It also seems that there's some confusion left regarding the various timing components of a <trial> and how they interrelate. I strongly recommend reading the "How to Control Trial Duration and Inter-Trial Intervals" topic in the documentation, which explains all of that.

RE: adopted Columbia Card Task

minin72704 — Tue, 11 Aug 2015 07:34:08 GMT

It may be the last two questions (i hope so). First one, revised BARTY.iqx works well, however, it seems exceeding 2s than expect. Specifically, I expected the totoal time is 3mins 16s (1s fixation+10s instruction+15s*12 balloon+5s feedback).

The second question is about cued go no go task (see attachment). I want that each trial was exact 4s (from <trial cue> to <trial target> to <trial iti>), so i add "/ posttrialpause = 3300-trial.target.elapsedtime-trial.cue.trialduration" in <trial.target>. However, it resulted in 8s more than expect, and the colored rectangle didn't disapear immediately after my response (in fact, I want it disapeared immediately after my response then wait for a while until next fixation appeared, just as default setting). Could you help me figure out how to resolve it? Thank you so much.

RE: adopted Columbia Card Task

minin72704 — Tue, 11 Aug 2015 06:52:08 GMT

Thanks for your help. It looks great.