Cumulative time wrong with responses?

Greetings.

I have run into an issue when trying to use the ‘Cummulative Time’ column in the output text file to calculate event onset timings/durations, particularly when there is a participant response. I am posting this in the SuperLab 4 support, however, this problem seems to exist in SuperLab 5 as well.

I have (as an example), events with the following assigned durations:
4000
500
4000
4000
4000
500
4000
500

If I run the experiment having the participant withhold responses, I can calculate the following event durations from the Cumulative Time:
4072
562
4064
4062
4061
562
4064
562

However, if I re-run the same experiment, this time having the participant respond as instructed, I calculate the following event durations:
1881
2755
2709
5424
1239
3396
1861
2774

These are clearly not correct. I have tried various ways to linearly combine the reaction times (1812, 0, 2643, 0, 1172, 0, 1794 - participant only responds to events # 1, 3, 5, 7) with the immediately above calculated durations, but I cannot find any logical way to retrieve the information I need.

Unless someone can correct me, the only way I can see to get accurate onset times is to extract them from an experiment run that does not involve any participant responses whatsoever. This should work fine for experiments in which the timings are consistent across all runs (e.g., no inter-subject randomization) but becomes nearly impossible for experiments that include inter-subject randomization. Is there a recommended way to get accurate onset timings for experiments that involve participant responses, as this seems like this should be a requirement for any fMRI stimulus presentation software?

Thanks in advance,

suzanne

Suzanne, I doubt very much that the cumulative time is wrong as SuperLab goes through exhaustive testing. I would need to look at your experiment to be able to interpret the data correctly.

Also, you mentioned fMRI stimulus presentation software. I would guess that you are trying to synchronize the trial presentations with the TR of the scanner. If this is correct, please see this page, “Synchronizing Trial Presentation With fMRI Scanner”.

I am not saying that the internal timing in superlab is wrong, as I have timed various stimulus durations using a stopwatch during runs with and without participant responses. I am also not trying to synchronize trials presentations with the TR. I am simply trying to be able to accurately extract event onset times from the superlab output text file.

Ultimately, my issue is with how the cummulative time is written out when there is a participant response, as what is written out in the ‘Cummulative Time’ column of the output file does not seem to reflect the actual timing of the experiment when a participant responds via a button press to a stimulus (we are using Lumina response pads). Is there some option within SuperLab that I am failing to select that will correct this? I have tried selecting and de-selecting the option to reset the RT clock for events, but this did not seem to have any effect on how the cummulative time was written out.

Again, any help would be appreciated, as this problem is currently affecting several studies at our research center.

suzanne

The cumulative time is not affected by any options within SuperLab. However, the reaction times can be affected by a setting called Reset RT timer. This is found in the Event Editor’s Input tab.

SuperLab always resets the RT timer on the onset of the first event in a trial. However, there are common situations where you would want to reset the timer yourself. An example is a trial where you present a cue for 500ms followed by a stimulus. If you do not explicitly reset the timer for the stimulus event, then all RTs are measured relative to the onset of the cue — not the desired outcome.