response time keyboard not saved

zyko

Dear all,

I made a modified version of SART. However, responses times from the keyboard press are not saved.

Actually, the first response time saved is set as "NA" and the following look randomly set between 0 and 2.

I know that's is due to the way I programmed the task but I can't find the mistake.

An idea?

Thanks,

Alex

SART_v4_NoProbes.osexp

Fab

Hi @zyko,

I had a look at your problem and I think I can see a few issues with the way your task is set up. I should however say that I'm no expert on co-routines (in fact, I've never used them), so there may be more issues to fix than those I'm commenting below and you'll need to check thoroughly the solutions I suggest.

(1) The first thing I noticed is that you actually set up your stimulus_presentation and mask stimuli to require a keyboard response. This make little sense, however, since the even you want to record is the response to the keyboard_response object.

The duration of your stimulus_presentation and mask should be 0.

(2) The next issue relates to the variables you inserted into the logger. There, you're mixing sequence-wise variables and keyboard response variables.

You're taking a response non specific to any object but you're taking the response time specific to the keyboard_response object. This is problematic in the light of the third issue described below.

(3) The set up of your co-routine makes it impossible for the task to register the responses you want to register.

For example, you start the keyboard_response object from the onset of the sequence, which is OK since you want to measure the RT from the onset of the stimulus_presentation. However, you also stop it at 0ms. Hence, the keyboard_response object could never actually register a response. Note that yo set the timeout property of the keyboard_response object to 1250ms, but I'm pretty sure that you're actually killing it right away because of the "End time" value you're using for that object in the routine declaration.

Again, I'm no expert on co-routines, but it seems to make little sense to set 0 ms as the start time of the logger, since there would be nothing to leg yet at that time. The logger should kick in once the events in your trial have occurred (especially the subject's response). I think you should set that value to a later point in time, though making sure that value is not greater than the value you set as the duration of the co-routine itself.

Proposed solutions

I had a quick go at modifying your task (again, I'm offering this as it is; you'll need to conduct checks and see whether this works for you).

(1) I set the duration of the stimulus_presentation and mask objects to 0.

(2) to show you that the above is corrected, sequence-wise and keyboard_response variables are equivalent, I inserted both in the logger:

(3) I modified the co-routine timings and inserted a wait sketchpad (duration of 1000ms here, but you could shorten that to something a lot shorter) to allow the keyboard_response object to run to its maximum duration (1250ms) and still have time to start the logger after that (if you execute the logger before a response has been registered, that response can obviously not appear in the logger). Note that I then had to modify the duration of the co-routine.

With these fixes, the log output now seems complete (space bar responses are all registered when before you had lots of "none", and RTs appear to be all registered too):

I attach my modifications below so that you can try it out:

SART_v4_NoProbes_Fab.osexp

Some last comments:

I noticed that you only log data for the test trials and not for the practice trials. I would recommend saving all the data and adding a loop variable where you can indicate whether trials are practice or test. You can then later on discard practice trials for your analysis, but that way you have the data if you ever need them (for example, to see which subject decided to repeat the practice). Data saved can always be ignored afterwards, Data not saved can never be recovered.

I assume you will run your task in the lab and not online (co-routines are not supported by OSWeb).

Hope this helps. Good luck!

Fabrice.