synchronized trigger and stim presentation

[Yisheng Xu]

Attached a program for testing the auditory and visual stimulus delay 
using MEG. We tested the delays on an Electa Neuromag VectorView system. 
The trigger signal was sent from a computer parallel port (address &H378) 
to the MEG trigger box with a customized DB37-DB25 adaptor cable. The 
auditory signal (a 100 ms 400 Hz pure tone) was sent from the computer 
sound card to the MEG system and measured by a microphone connected to an 
ER-5A ear insert. The visual signal was projected by a Panasonic PT-D7700U 
DLP projector to a translucent screen. The projector was connected to the 
computer VGA port. The refresh rate of the graphic card (Intel 910GL) was 
set to 60 Hz. A photodiode was fixed with a sucking disk to the center of 
the screen to measured the luminance of a repeated 100-ms flash sginal 
presented by the program. Both the microphone and the photodiode (with 
internal preamplifiers) were connected to the miscellaneous channels of 
MEG. In this way, we can accurately measure the auditory and visual delays 
relative to the trigger signal in an accuracy of 0.1 millisecond.

E-Prime FAQ recommends to use StimDisplay.OnsetSignalData to synchronize 
the trigger with the stimuli. Ideally, we may use a slide object to send 
auditory and visual stimuli together with the parallel trigger. The major 
problem of this method is the visual stimuli is not displayed at the 
actual onset of the slide object although E-Prime has a function to 
synchronize the image with the vertical blank of display. In this simple 
setup, the measured audiotory delay is 3-4 ms depending on the sound card; 
The measured visual delay is 20-30 ms depending on the graphic card and 
refresh rate. Using our hardware setup, for an unknown reason (maybe due 
to the native refresh rate of the DLP projector), refresh rates other than 
60 Hz (e.g., 75 Hz, 85 Hz) will generate an inter-trial jitter of the 
visual delay, i.e., the delay of the visual signal varies among trials. As 
a result, the averaged luminance function will show a sloped onset and 
offset and the refresh cycles will be smeared. For the timing requirment 
of EEG/MEG, a fixed delay is necessary. At a 60 Hz refresh rate, the onset 
and offset of the measured luminance function are very sharp. The refresh 
cycles can also be clearly identified. On some slow computers, interrun 
timing variations were also observed besides the intertrial variation, 
i.e., we may observe different averaged delays in different runs even 
though no intertrial variation at 60 Hz.

Because we're doing some multisensory integration study, the 
sychronization between auditory, visual and trigger signals is critical. 
The attached program using the DisplayDevice.WaitForVerticalBlank method 
and Sleep statement to adjust the timing delay. We achieved a measured 
auditory delay of -0.1 ms and a measured visual delay of 0.5 ms. The 
reason that the values are not exactly zeros is because the operating 
system (DirecX) only allow for integral millisecond increments (for the 
Sleep statement). The interrun variation is less than 0.1 ms.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: C:\Documents and Settings\All Users\Documents\Shared Experiments\Trigger Tests\trigger2stim_delay.es
Type: application/octet-stream
Size: 21783 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/eprime/attachments/20060620/b592d77b/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: C:\Documents and Settings\All Users\Documents\Shared Experiments\Trigger Tests\400Hz_100ms_testtone.wav
Type: audio/wav
Size: 17684 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/eprime/attachments/20060620/b592d77b/attachment.bin>