Dear Donald & Fanny,

I would like to add my 5 cents to Donald's excellent suggestions.

In randomized ER designs, there are 2 schools concerning the choice of
ISI, people that use very long SOAs as proposed by Donald, and others
that use randomized short ones + null events, as Rik Henson (2004, ref
below), or as pointed out by Karl Friston in his 1999 paper (ref
below).

The choice of ISI depends on your needs & hypotheses, the number of
conditions you want to test (ie probability of occurrence), whether
you aim at characterizing differential effects (in which case short
SOAs are more efficient), whether you aim at characterizing the evoked
response itself using FIR models (in which case long SOA as suggested
by Donald are preferred), etc

As pointed out by Friston et al, short SOAs are more efficient and
maximize the efficiency of your model, provided your events are
correctly randomized. The smaller the SOA, the greater the number of
events that can be presented, and therefore the more efficient the
design is. Also, short SOAs ensure "Rapid presentation rates allow for
the maintenance of a particular cognitive or attentional set, decrease
the latitude that the subject has for engaging alternative strategies,
or incidental processing"

A nice compromise between both views, which has been addressed by
Haberg et al 2001 (ref below), is to use long-tailed probability
distributions such as geometric or exponential distributions with a
long decay. These distributions have the advantage of having many
short events + a few long ones that can be seen as null
events.Personally, I use exponential or poisson distributions, with a
mean of ~4-5 seconds and a range of 1 to ~8 seconds.

All the best

Swann

Stochastic designs in event-related fMRI.
Friston KJ, Zarahn E, Josephs O, Henson RN, Dale AM.
Neuroimage. 1999 Nov;10(5):607-19.

Improved detection of event-related functional MRI signals using
probability functions.
Hagberg GE, Zito G, Patria F, Sanes JN.
Neuroimage. 2001 Nov;14(5):1193-205.

Henson RNA (2004) Analysis of fMRI time-series: linear time invariant
models, event-related fMRI and optimal experimental
design. In Human Brain Function, 2nd edn, Frackowiak RS, Friston
KJ, Frith CD et al. (eds). Elsevier, London, pp 793–822

--
Swann Pichon, PhD
Laboratory for Behavioral Neurology and Imaging of Cognition
Department of Neuroscience, University Medical Center
1 rue Michel-Servet, 1211 GENEVA 4, Switzerland
Tel: +41 (0)22 379 5979
Fax: +41 (0)22 379 5402
Gsm: +33 (0)6 26 43 83 61
http://labnic.unige.ch/




Le 24 juin 2011 01:43, MCLAREN, Donald <[log in to unmask]> a écrit :
>
> Fanny,
>
> You have a 10 second trial. The 1s video is independent of this
> initial statement since its entirely enclosed.
>
> You will want to model the thermal stimuli, the video, and the rating
> period as 3 different conditions. Thus, you will end up with 12
> conditions. I would recommend that you have 30 of each condition (120
> 10s trials).
>
> Although, you don't need fixation, you will want more than what you
> have suggested for several reasons: (1) having only short 1 second
> interval will be stressful for the subject; (2) you are wanting to be
> able to accurately model the thermal, video, and rating periods, so
> more gap will allow the response to come back to baseline before the
> next trial.
>
> I'd personally aim for inter-trial-intervals of 5-9 seconds (1 sec.
> increments would be fine). Begin the experiment with 21 seconds of
> fixation (you'll discard the first 2 TRs and then have 5 more to get
> an estimate of the baseline values. Also, make sure you have some time
> after the trial to measure the last trial response coming back down to
> baseline (15 seconds)
>
> The minimum ITI ever needs to 1-2 seconds, any shorter and the events
> won't be separable and you can non-linear effects.
>
> Best Regards, Donald McLaren
> =================
> D.G. McLaren, Ph.D.
> Postdoctoral Research Fellow, GRECC, Bedford VA
> Research Fellow, Department of Neurology, Massachusetts General Hospital and
> Harvard Medical School
> Office: (773) 406-2464
> =====================
> This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED
> HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is
> intended only for the use of the individual or entity named above. If the
> reader of the e-mail is not the intended recipient or the employee or agent
> responsible for delivering it to the intended recipient, you are hereby
> notified that you are in possession of confidential and privileged
> information. Any unauthorized use, disclosure, copying or the taking of any
> action in reliance on the contents of this information is strictly
> prohibited and may be unlawful. If you have received this e-mail
> unintentionally, please immediately notify the sender via telephone at (773)
> 406-2464 or email.
>
>
>
>
> 2011/6/23 Fanny Eugène <[log in to unmask]>:
> > Dear SPMers,
> >
> > I need some advice on timing parameters for an event-related design. Each of my trials lasts 10 seconds, with a 1-second video clip presented 4 s. after the start of the trial. I’m only interested in brain activation during the clip and how it varies across 4 conditions (4 types of video clips). I plan to present a fixation cross of varying durations between trials.
> >
> > From what I understand, since I plan to contrast activation across conditions and I’m not interested in contrasting each condition with a baseline (null events), the duration of my fixation cross does not need to be very long, but only needs to shift the timing of my events so that they do not always start at the beginning of a TR. Is this right?
> >
> > With a 3 sec TR, would it be ok for me to use a fixation cross with durations that vary randomly from 500ms to 2500ms in steps of 500ms?
> >
> > Any advice would be very helpful, thanks!
> >
> > Fanny
> >