Can't reduce onset delay

[Chen G]

Hello,
I have an experiment with a total duration error of 500-1500 ms for a 3:14 min expected duration. The experiment is intended for fMRI so I need accurate duration for stimuli as well as total experiment time (I set the objects in the script to Event timing mode).
I have tried to reduce it many ways (adjusting PreRelease, Frame size, Procedure.GeneratePreRun, updating monitor driver) but couldn't find anything that works.
The paradigm is as follows:
- alternating blocks of stimuli and fixation.
- stimulus block: 4 trials of the following structure:
(1) Slide object: bmp image (tried also jpg) with a slide text on top of it (a fixation cross presented simultaneously with the image): 2700 ms
(2) Text display (fixation cross): 300 ms
- fixation block: Text display (fixation cross): 10 sec.

The experiment advisor and the edat file suggest the following:
- Slide object: in every block the first object has OnsetDelay of 0 or -1, and the other three all have onset delays of 16-34 ms.
- The text display that follows each of the slide objects always has OnsetDelay of 0 or -1.
- The text display in the fixation block always has onset delay of 16-17 ms.

* I suspect that the source of the problem might be in the relatively large images (1 MB as bmp, 80 KB as jpg), but I can't do much about that (I use a similar script for another experiment with smaller images and I don't have timing issues there).

Interestingly, I tested the script with E-Run Test and all these onset delays vanished (total duration error for experiment: 19 ms).

I am using E-Prime 2.0.10.356 on a Windows 7 desktop computer with an LCD display with resolution of 1920x1080 and refresh rate of 60Hz.
Any ideas what to do?

Thanks,
Chen

[David McFarlane]

Chen,

First and foremost, for fMRI, set the timing mode of all your objects to
Cumulative, instead of Event! To understand why, work through the
"Critical Timing" chapter of the User's Guide that came with E-Prime, or
search through this Google Group where this has been explained several
times before (or take my online course).

Your OnsetDelay values actually look reasonable to me. The 34 ms delay
may have more to do with the high resolution setting of your display
than the size of your image files (for our fMRI we only use 1024 x 768),
while the other delays may be ordinary delays due to waiting for the
vertical blank. You should find all of this also explained in the
sources above.

Interesting, indeed, that you get a vastly different total duration
error when using E-Run Test. I too have noticed that E-Run Test does
things a little differently than running through E-Run (E-Run Test also
seems to not even create an .ebs2 file), and I have no explanation.

Good luck!

---------------
David McFarlane
E-Prime training online:
http://psychology.msu.edu/Workshops_Courses/eprime.aspx
Twitter: @EPrimeMaster (https://twitter.com/EPrimeMaster)

[Chen G]

Thank you, David.
That's good advice!

Chen