We are using VC++ and OpenGL in our lab to program the visual stimuli so there is little concern about synchronization. 

To get back to your questions (I can only try to answer the first two):

Using PTB on a Laptop is problematic if you do not have a dedicated graphics card. Most (!) Laptops only have onboard GCs, which are low quality and will cause sync failures. This is more often the case for Windows systems. Here, especially Windows 7/10 are cumbersome as you can not turn of the Window manager anymore. In addition, LCDs are generally less timing friendly than CRTs - to be very polite. The best setup would be to use a Linux System with a CRT monitor (for the full "oculomotor nerd setup" take a look at Carpenter's book Vision Research). Since this is again a switch-the-system answer, a solution for you setup would be to add a photodiode to the screen for which you record the output as a synchronized stream to your eye data. This gives you precise timing information of your stimuli.

There are also very good resources from PTB on that topic:

• Hardware & GC requirements: http://psychtoolbox.org/requirements
• Sync failures:

- http://docs.psychtoolbox.org/SyncTrouble

- http://docs.psychtoolbox.org/VBLSyncTest

• Simple flip timing test: http://docs.psychtoolbox.org/FlipTest
• General Q&A: http://github.com/Psychtoolbox-3/Psychtoolbox-3/wiki/FAQ
• Timestamps explained: http://github.com/Psychtoolbox-3/Psychtoolbox-3/wiki/FAQ%3A-Explanation-of-Flip-Timestamps
• PTB performance tuning: http://github.com/Psychtoolbox-3/Psychtoolbox-3/wiki/FAQ%3A-Performance-Tuning
• Problems using Win Vista+: http://github.com/Psychtoolbox-3/Psychtoolbox-3/wiki/FAQ%3A-Vista-and-Windows-7

Another useful resource for writing experiments in PTB is Peter Scarfe:

http://peterscarfe.com/ptbtutorials.html

And finally the work by M. Kleiner is very helpful (see pub).

For the second question, I would just advise to be aware that using a visual example is easier (as it instantaneously gives you a template to look for) but will prime a location. As you will most likely start in the center of the screen anyway, that should not be a big problem - just keep it in mind.

Greetings, David

Article Visual stimulus timing precision in Psychtoolbox-3: tests, p...

Article Vision Research: a Practical Guide to Laboratory Methods, R....

Rui,

Is there any guideline that you follow for selection of images as visual stimuli? Did you specify the number or duration of display based on certain criteria?

David,

Thank you for your detailed answer as usual.

Regarding the photo-diode suggestion, can you elaborate as I didn't get how adding it will help with accurate synchronization.

During the visual search experiment, did you re-calibrate after every scene displayed or how did you ensure the accuracy of calibration throughout the experiment

Hey Jihad,

1. Get a photodiode. Plug that on a corner of your screen. Draw some small rectangle together with your stimulus at that location -> Will result in a luminance change detectable by the diode. Record the signal from the diode synchronized with the eye data and you have the rather exact timing of the stimulus (given that there might be some small difference when the stimulus and the rectangle actually appear on the screen).

2. If you use a head/chinrest, calibration should be stable enough for normal measurement durations. For long experiments or without a fixed subject, recalibration is the way to go, but will make your experiments longer. So you have to decide on the speed-accuracy trade-off.

Greetings, David

Hi Jihad,

The selection of images really depend on your research question and design. If you want to study scene perception for instance, you'll gather a good sample of images that represent the visual world. As for the display duration, anything shorter than 200ms is usually considered being safe to avoid eye movement. 

Rui Ni

Hi Rui

Regarding the the display duration, isnt the purpose is to record the eye movements while the subject is performing one of the visual tasks? thus 200 ms is quite short as many papers vary between 10 s as minimum and no time limits as maximum.

Hi Jihad,

For visual psychophysics studies, shorter display duration is a common practice (

Thank you for the help :)