How to implement ordered stimulus repetition in E-Prime

[Valerie]

Hi,

For my fMRI adaptation experiment, I need to encode ordered
repetitions. That is, I have an ordered list of stimuli that will be
presented, but within this list (each of the) 12 items will be
repeated three times (with a variable number of intervening trials). I
was wondering, is there a way to make use of the ID’s that the
different stimuli have to implement this repetition in E-Prime? For
example, by writing some script which recalls an ID or something?
(I.e. [present 1, 2, 3, 1, ..]). I will need to make 125 lists, so it
would really save time if there were a way to implement this.

Thanks!


Valerie

[liwenna]

Hi Valerie,

I think we are a bit in the dark on your exact needs still. If you'd
write a script within e-prime to order the triallist each run of e-
prime will give the exact same order, which isnt' what you want I
think cause if I understand correctly you need 125 different 'orders'.
Right? And these orders need to be known beforehand and can't be
created 'on the fly', am I right to think so?

In that case I think your best shot is to create 125 lists and store
them in 125 textfiles that are loaded into the triallist by e-prime
based on for instance subjectnumber.

If the randomisation can be made at random for each subject during the
running of e-prime... you could consider using three different lists
that are run subsequently each of them containing your 12 key-trials
as well as a number of fillertrials and randomise each list: this way
you can't control the number of fillertrials inbetween keytrials and
also you can't prevent keytrials to be presented subsequently but you
do know that each of the 12 keytrials will have been presented once
before any of them is repeated and you also have a random number of
intervening fillertrials as well as a new random order on each run of
e-prime (i.e. each subject).

Just two cents, tell a bit more about your exact needs if the above
doesn't help at all (which seems likely :p)

Best,

liw

[David McFarlane]

>I think we are a bit in the dark on your exact needs still. If you'd
>write a script within e-prime to order the triallist each run of e-
>prime will give the exact same order, which isnt' what you want I
>think cause if I understand correctly you need 125 different 'orders'.
>Right? And these orders need to be known beforehand and can't be
>created 'on the fly', am I right to think so?

I didn't get what Valerie wanted either, liw, but your response rings
a bell for me. In our fMRI studies we often use a limited number of
fixed pre-randomized orders so as to avoid the problem of
multicollinearity during analysis of the functional brain images --
for correlation, deconvolution, or general linear modelling analysis
to work we cannot allow any sequence to be a linear combination of
any others, so we generally cannot leave this to chance by
randomizing sequences "on the fly" (this is explained very nicely in
the AFNI documentation from NIMH).

>In that case I think your best shot is to create 125 lists and store
>them in 125 textfiles that are loaded into the triallist by e-prime
>based on for instance subjectnumber.

You might also do this all directly in E-Studio with no external text
files and no script. You would have to make a master List with one
row (level) for each possible sequence (in this case, perhaps 125
rows). You would then set its selection Order to Counterbalance, and
Order By to Subject (or perhaps to Session). Counterbalance will
select and run one *and only one* row from the master List, based on
Subject (or Session). (This is explained somewhere in the EP docs,
or just search the EP Group with the term "Counterbalance" to see
where I recently explained this.) Back in your master List, you
could have each row call a customized Procedure for one of the 125
random orders, or have each row use the same Procedure but a
different nested List for the random order, etc.

Don't know if this addresses Valerie's question, but it does address
an fMRI issue that crops up from time to time, so I am adding this to my FAQ.

-- David McFarlane, Professional Faultfinder