User`s guide

E-Prime User’s Guide
Chapter 3: Critical Timing
Page 111
33.24. If matching the Probe to Mask times with the Probe Duration times, the Probe Durations
are 10ms less, as expected.
An additional check is needed to determine that the durations in each condition were as
expected. A pivot table can be generated in Excel to show the results broken down by
ProbeDuration.
ProbeDuration Mean Probe-Mask S.D. Probe-Mask
7 16.88 0.33
23 33.25 0.43
40 49.75 0.43
The mean for each Probe to Mask duration is 10ms more than the ProbeDuration, and is in close
agreement with the expected duration (i.e., observed 33.25ms; expected 33.24). Importantly, the
standard deviation for each condition was below 0.5ms. A variation of 1ms is expected due to
rounding effects when logging the data in integer numbers of milliseconds. The observed timing
characteristics can now be reported; fixation (498.7ms, standard deviation 0.47), with probe
durations of 16.88, 33.25, and 49.75ms, with a standard deviation of less than half a millisecond.
If the timings are not in close agreement as above, check to be sure that there were no errors in
specifying the ProbeDuration as multiples of the refresh rate, that the Event timing mode is in
use, and that PreRelease is set correctly
13
.
Timing Paradigm 4. Cumulative timing of a repeating sequence of events
Cumulative timing paradigms occur when seeking to present a sequence or stream of stimuli for a
long period of time and carefully control the rate of presentation. You may also be maintaining
that rate with an external recording device. A behavioral example might be a continuous
performance task presenting stimuli every second for twenty minutes. A second example of such
experiments is one in which you are concurrently recording data and need to maintain precise
synchronization with the other device (e.g., brain wave monitoring equipment or functional
imaging equipment).
There are four important differences between an experiment applying critical timing to a short
sequence of events (Timing Paradigm 2) and one presenting a repeating sequence (Timing
Paradigm 4). First, the scope of critical timing involves more objects and includes stimulus
selection (see Figure 20). Second, because we maintain time to a fixed standard, all of the
durations must be fixed (i.e., cannot vary trial to trial) and subject responses cannot alter the
overall timing sequence (e.g., terminate the trial upon response). Third, the E-Prime clock may
have to be rescaled to match the timing of the external clock. The fourth difference is that there
may be some synchronization code so that one button starts both the behavioral computer and
the recording computer.
For illustration, we will adapt the previous experiment to operate in a Cumulative timing mode.
The first thing to note is that in Cumulative timing mode, the number of events under critical
timing control is much larger than in Timing Paradigm 2. This is because we need a loop that
repetitively selects the stimulus (TrialList) and runs the Procedure (TrialProc), including the
Fixation, Probe, Mask and Feedback.
13
Another common source of error is missed refreshes.