Startle probe complex randomization

[cblaison]

Hi all, I got an answer from PST, for the one who are interested in..
See googledocs for the modified file:
https://docs.google.com/leaf?id=0Bw6aNdKBxq4tM2JlOGE4N2ItNmY1Ni00MDY3LWJmMDctMzY2YTg2MWNmYjdl&sort=name&layout=list&num=50
and:
https://docs.google.com/leaf?id=0Bw6aNdKBxq4tNGE5ODU1ZTYtMWJiMC00YjNkLTg0NTAtMWExZjZlOTdjZDMy&sort=name&layout=list&num=50

PST answer:
Hi!

I attached the prog to be modified. It is a startle probe
prog, where angry faces are shown during 6 sec and sometimes bursts of
white noise are triggered. The Intertrial Interval (ITI) between faces
lasts between 14 and 21 sec. There are 4 different angry faces. The
probe starts 4, 4.5, or 5 sec after onset of stim (face or ITI). In
all, faces are presented 12 times, separated by 11 ITI + 1 ITI at the
end of the sequence. 3 Bursts have to appear during the ITI (which ITI
is random). 4 bursts have to appear randomly during the face
presentations BUT once appeared during one face, it never has to
appear again paired with this face. Thus e-prime has to recall somehow
which face was paired with the burst and discard this candidate from
future face-probe pairing. I am not an in-line expert, so could please
someone help me? Thank you very much!

cheers,
christophe.

>From Alison Lauer on 2/14/2011 5:11:00 PM
Conditioning2MOD2.es2
Dear Christophe,

Thank you for attaching your experiment files. I will be happy to
help.

First, I recommend restructuring your experiment to better reflect
your needs.
- It is best to rename your objects to meet standard E-Prime naming
conventions. This will help to keep the experiment clear to yourself
and others. The name "proc" in E-Prime represents procedure, so the
convention is to name your lists BlockList, TrialList, etc. and to
name your procedures BlockProc, TrialProc, etc.
- The sound choice (between 9 silent trials and 3 startle trials) can
be moved from probchoice to FaceList. Both probchoice and
probchoiceITI can be eliminated completely since their functions are
fulfilled by FaceList and ITIList respectively.
- Instead of nesting startlechoice under probchoice, nest it at a much
higher level, like BlockList. We will be using a bit of InLine script
to access this list, so we only need to put it at the block level for
reference.
- Add two new attributes to facechoice: one for the trigger code (the
answer to request 22022) and one for the startle sound.

Next, you can use an InLine object to get the three sounds and fill
FaceList with their values. You may want to take a look at one of our
samples for a similar (though more complicated method: SAMPLE:No
Repeats on Consecutive Trials (C.... Here is the script to use:
Dim arrSounds(3) As String

Dim i As Integer
For i = 1 To startlechoice.Size
   arrSounds(i) = startlechoice.GetAttrib(i, "startlechoix")
Next 'i

RandomizeArray arrSounds, 1, startlechoice.Size

For i = 1 To startlechoice.Size
   facechoice.SetAttrib i, "FaceSound", arrSounds(i)
Next 'i

facechoice.Reset

As for request 22022, you can use an attribute value to vary the
signal that is sent to the port. Create an attribute named FaceCode on
facechoice, then add the following script in an InLine on FaceProc:
WritePort &H378, c.GetAttrib("FaceCode")

Finally, keep in mind that your description may not be possible with
the current number of stimuli. If there are 3 sounds and 4 faces, one
of the sounds would have to be repeated for a face. Do you mean that
each face must have a unique sound? Or do you mean for one of the
faces to have a silent probe? The attached version of the experiment
has the fourth face disabled (by setting its Weight to zero) in order
to use a unique sound for each face. If you have something else in
mind, please describe it in as much detail as you can.

I hope this helps. Please let me know if you have any more questions.

Sincerely,
Alison Lauer
Technical Consultant

>From christophe blaison on 2/15/2011 8:59:00 AM
Pretest table.docx
Thank you alison, I really appreciate your quick reply!

I was perhaps not clear enough in the description: The solution you
propose is on the right track but not perfectly what I would want the
prog to do. I try it again (see also attached figure)

The startle probes must be randomly sampled from the pool of three
startle probes (4s, 4.5s, 5s after stim onset) and have to appear only
on one (of three) angry face presentation occasion. In fact, the four
angry faces are presented randomly in 3 presentation blocks: each face
must appear only once in each block. Thus there are four occasions
where the startle is associated with angry faces, and startle must not
be associated two times with the same face. (see attached figure).
In addition the randomly chosen startle probe have to appear only 2
times whithin the ITIs (see attached figure).

Am I clear enough now? Thanks!
christophe.

>From christophe blaison on 2/17/2011 3:18:00 AM
Hi Alison, I know you are certainly as busy as me, but I wonder if you
had some time to resolve my problem, or if you need some additional
information.
Cheers,
Christophe.

>From christophe blaison on 2/20/2011 7:02:00 AM
Hi Alison, thank you for closing my other request. We are a bit in a
hurry here (I am aware that they are probably other people in our
case), thus I would be very grateful to you if could send us a
solution to the problem soon! Thank you :)

>From Alison Lauer on 2/22/2011 4:57:00 PM
Conditioning2MOD3.es2
Christophe,

Thank you for explaining your experiment in more detail. I apologize
for the delay in my response. For your reference, please note that
when you reply to the request before I do, your request is
automatically sent to the end of the queue, thus extending my response
time further. Please feel free to ask if you have any questions about
our support site system.

I have determined how to restructure your experiment for your needs.
Since you require each block to randomly present faces 1-4 while
playing the startle only once for each sound throughout three blocks,
you must keep the startle/silence selection independent for each face.
TrialList should select one of four faces for each trial. Face1List,
Face2List, etc. should be nested under TrialList, and each one should
have the same three levels: one for startle (in which StartleList is
nested), and two for silence (or one for silence with Weight set to
2). The Face lists will not be reset between blocks, so only one
instance of each face will be associated with a startle. ITI durations
can be selected from ITIchoice list nested under TrialList, with
ITISoundList nested under that. ITISoundList can use the Weight method
to ensure that two levels have a startle for ITI, and the remaining
ten have silence. For simplicity you can place AF and ITI onto the
same procedure instead of running them from different lists.

I have attached a modified version of your experiment. Please take a
look at this version and let me know if you have any questions. I will
be happy to answer.

Alison Lauer

>From christophe blaison on 2/25/2011 1:26:00 PM
conditioning christophe.zip
Hi Alison, thank you very much for your help, it works like a charm
and we learn a lot from you.
Would you be so kind to make two other useful modifications? We would
be very grateful to you. We build on your example for programming the
complete experiment (see attached). We now have a preconditioning
block (the one you worked on, it contains just startle probes), a
conditioning block with electric shocks and startle probes and a
postconditioning block (just startle probes).

1. Electric shocks are associated with two of the four angry faces (we
send a 128 code through the parallel port to trigger the shocker). We
would like to render this pairing random accross participants. For now
we created different eprime files containing different pairing. But
this solution is not satisfying.

2. A crucial feature for us in order to analyse the data is to be able
to identify during which angry face a startle probe has been sent.
Thus it would be useful to send a different startle probe marker for
each angry face. You already did a similar kind of things for the
presentation of the angry faces..but this seems to us a bit more
difficult.

We thank you very much, and we are eager to see your solutions!
Cheers,
christophe.

On Feb 13, 10:52 pm, cblaison <hessl... at googlemail.com> wrote:
> Hi all!
>
> I attached here:
>
> https://docs.google.com/leaf?id=0Bw6aNdKBxq4tM2JlOGE4N2ItNmY1Ni00MDY3...
>
> a copy from the prog I would like to modify. It is a startle probe
> prog, where angry faces are shown during 6 sec and sometimes bursts of
> white noise are triggered. The Intertrial Interval (ITI) between faces
> lasts between 14 and 21 sec. There are 4 different angry faces. The
> probe starts 4, 4.5, or 5 sec after onset of stim (face or ITI). In
> all, faces are presented 12 times, separated by 11 ITI + 1 ITI at the
> end of the sequence. 3 Bursts have to appear during the ITI (which ITI
> is random). 4 bursts have to appear randomly during the face
> presentations BUT once appeared during one face, it never has to
> appear again paired with this face. Thus e-prime has to recall somehow
> which face was paired with the burst and discard this candidate from
> future face-probe pairing. I am not an in-line expert, so could please
> someone help me? Thank you very much!
>
> cheers,
> christophe.

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.