Hi there, I’ve noticed a few threads about timing in SL, and I have my own question regarding this.
We have an experiment that involves a one-paragraph story being displayed (as text) and left on screen for 10s. It is synched to an MRI scanner (via an RB-834) which pulses once every 2s. In the data file we get some consistently incorrect RT recording, such that SL records the first pulse incorrectly after about 1948ms, and the subsequent 4 pulses correctly after about 2000ms (± 2ms or so).
This pattern occurs regardless of the stimulus type, or how long it is to be displayed for. In another case, we’re presenting a crosshair (+) for 6s - the first pulse is recorded at 1965ms or so, and the following 2 at 2000ms (± 2ms).
My feeling is that the time taken to draw the corresponding text to the screen is not counting towards the total RT, which can make timing quite difficult (especially since we’d like to synch stimulus onset with scanner TRs). Is there any way to either make the drawing time non-existent, or to have the RT timer start immediately at the beginning of the trial (rather than once the text appears)?
Without seeing your experiemnt, I’d say that your guess regarding the RT not including the drawing time is correct. For visual events (text, pictures), the RT is always measured from the onset of the event, and the timer can be reset so that it can start from the onset of the 2nd or 3rd event in the trial, e.g. if you have a cue followed by stimulus, you typically do not want RTs measured from the onset of the cue, so one must reset the timer in the second (stimulus) event.
The point is, for visual stimuli it’s always measured from the onset – after the image appears.
Also, the missing 52ms (2000 minus 1948) are not just the stimulus’ drawing time. If that event happens to be the first one in the trial, the 52ms also include the necessary internal housekeeping that SuperLab must do as it moves from one trial to the other.
Is there any way to either make the drawing time non-existent, or to have the RT timer start immediately at the beginning of the trial (rather than once the text appears)?
One possibilty is to create a “dummy” event that precedes your stimulus. This will cause the timer to be reset for that particular trial. There will be some delay between the dummy event and the stimulus, but my hunch is that since it’s no longer the first event in the trial, it will be a lot less than 52ms.