How to measure trigger-stimulus delay ?Due to electronics, you will have a delay of few milliseconds between the time at which your triggers are being sent to the MEG acquisition computer and the time at which the stimuli will actually be displayed to the subject. No worry, this is normal BUT you need to know, quantify and take into account these delays (otherwise what's the deal doing MEG if it is to lose temporal resolution?!) Here is how you should proceed BEFORE your first pilot recording:
Set up the devices
If you use the photocell, place the sucking cup on the visual display you wish to measure (the one that a trigger should specify). If you use the microphone, place an earplug into the transparent tubing on one side and the mic on the other side. Then use the earplug as you usually do to connect to the Etymotic system. The cabling from the photocell and the microphone should be passed through the small opening into the stimulus cabinet outside of the shielded room (just like the optic lines of the forp button presses system). In the stimulus cabinet, plug the cables into their respective devices. Connect a BNC cable from the boxes into the circular metal piece with BNC connections in the inside of the stimulus cabinet. Then, take the longest BNC cables you can find and plug them from the circular metal piece with BNC connections this time from the outside of the stimulus cabinet (respect the order and trace back which line is recording what) all the way to open MISC lines through the device on the shelf above the eye-tracking PC. The first two lines are connected to the Eye Tracker, you can use any of the other lines available. Acquire sample data
* on the MISC channels, the time at which sounds or images are sent to the subject
* on the STI channels, the time at which the triggers corresponding to your stimuli are being received by the acquisition MEG
Analyse your sample data
maxfilter -f myfile_triggers.fif -nosss
This will produce a myfile_triggers_nosss.fif This is the file you want to use to measure the delays.
Load the Settings corresponding to "triggercheck.setup" or something sounding like this (you may want to create your own as different MISC channels may be used). Ask for help if you need to!
Load your file by using Load Disk File
You should measure the difference between your trigger onset (rising onset edge) and the stimulus onset (rising onset edge). This is your delay.
One measure is NOT sufficient. You need to measure this delay at different times during the recording, insure that all stimuli (even if they are all visual or all auditory) have in fact the same delay (due to load time, memory issue etc in the PC stim, you could have different delays for different stimuli). Double-check the consistency of your delays Once you have a sufficient sample of data points, do your stats! And make sure the variance is very low - best would be 1 to 2 ms of jitters across trials at most.
Correct your data accordingly You can now use these delays in your acquisition setup and/or correct offline in your preferred analysis software.
|