We have been starting to use EPrime and have found that it is possible to
get precise timing, but you certainly have to work at it! To avoid
accumulating delays, caused by successive stimuli being presented for
slightly more time than requested, EPrime has a "cumulative" timing mode,
which basically corrects for such delays by subtracting a small amount of
time from the following stimulus. In a blocked design, this sort of timing
change (typically a few milliseconds) is insignificant. In an event related
design it is potentially more so, although a few milliseconds error with a
TR of 2 or 3 seconds is not very much a problem. Small timing errors of
this sort are unavoidable unless the requested stimulus duration is an even
multiple of the screen frame refresh.
We have achieved single frame presentations of visual stimuli with a screen
refresh rate of 75 frames/second (verified with an oscilloscope). We have
also found EPrime's logged timing values to be an accurate indicator of
actual stimulus timing.
There are still a few bugs and quirks in EPrime, and it certainly can be a
bit of work to get an experiment up and running with accurate timing, but
it is possible, and once you learn the tricks it becomes easier. The
advantage is that EPrime is very flexible and doesn't require scripted code
(although it can be scripted).
Tom Johnstone
W.M. Keck Laboratory for Functional Brain Imaging and Behavior
University of Wisconsin-Madison
[log in to unmask]
+1 608 262 9230
At 04:18 AM 10/30/01, Chris Rorden wrote:
>Massimo-
>
>I use E-Prime in my student labs. It is a good teaching tool. However, I
>would offer extreme caution in using this software for research. There are
>several things you should consider:
> 1.) Stimulus timing may not be precise. Using an oscilloscope, we found
>that E-Prime could not present stimuli faster than 3 frames.
> 2.) Left to it's own devices, E-Prime will always lose a bit of time per
>trial. If you request it to wait 300 ms, it will wait AT LEAST 300ms, but
>sometimes it will be considerably longer. The data log will record these
>errors. I found that for a series of trials that were supposed to take 12
>sec, the actual time varied between 12.1 and 13.5 seconds (+10%). This
>would be a huge problem with block designs, and one that would at least
>need to be considered with for event-related designs. It may be possible to
>write some inline code that will help resynchronise the timing (e.g. at the
>end of each trial, wait until a certain time has elapsed), but if you
>simply add a standard E-Prime wait, it will not correct for previous delays
>- it always waits for AT LEAST the time you specified.
>
>All this was done ove the summer - PST may have improved things with the
>latest beta release. However, before using this software with any
>experiment, I strongly advise you to carefully look at the timing issues.
>
>-chris
>
>At 10:39 30/10/2001 +0100, Massimo Danna wrote:
>>Dear colleagues working with SPM,
>>
>>Our names are Massimo Danna and Pasquale Gagliano. We are two engineers of
>>the group of San Raffaele Hospital in Milan who like neuro researches. We
>>work with Ing Paola Scifo.
>>Here we are using a 5.7 Signa GE 1.5 Tesla Scanner for functional Magnetic
>>Resonance and we want to realaze event-related protocols so we need a
>>exact synchronization between tomograph and PC/MAC. Sorry but we did
>>find any e-mail about this in the email list so we have some question
>>about it:
>>
>>1. Which kind of system is there to realize event-related protocol?
>>
>>We have thought to buy 2 cards I/O (PCI-1200 National Insruments and
>>SC-2071 National Insruments plus an exernal home-device containing a
>>flip-flop D:
>>2. Are these cards good or not?
>>3. Do somebody know if this kind of cards are useful and with which
>>software of stimuli?
>>
>>Moreover we have thought to buy E-Prime:
>>4. Is this software good or not?
>>5. Are these cards compatible with E-Prime?
>>
>>Sorry to trouble you but this is the first time that we use GE Scanner
>>interfaced with a PC/MAC for event-related protocol.
>>Thank you very much in advance for any help.
>>
>>Massimo and Pasquale
>
>__________________________________________
>Dr. Chris Rorden
>Department of Psychology
>University of Nottingham
>Nottingham NG7 2RD, UK
>Tel: +44 [0]115-951-5294
>Fax: +44 [0]115-951-5324
>http://www.psychology.nottingham.ac.uk/staff/cr1/
|