Presentation pauses and returns to beginning in middle of experiment

stoney

Hi, I'm having a weird issue where, in the middle of an experiment, Open Sesame suddenly switches to showing the editor and minimises the experimental presentation window. When I maximise the experiment window, it shows the participant instruction screen, i.e. the experiment has gone back to the beginning. The Python console says the experiment has been paused.

I thought perhaps the participant accidentally hit Escape, but usually this would trigger the presentation of "The experiment has been paused. Press the spacebar to continue or Q to quit." and this wasn't the case. I also thought maybe I ran the experiment in a browser instead of fullscreen, but I can't reproduce the issue by doing that deliberately. It's happened across multiple experiments with different set-ups which were not all copied from each other (i.e. it might be a general problem).

Just wondering if someone else has had the same issue and what the cause was. Maybe a Windows notification forced active windows to minimise? (I've now turned all notifications off). System specs and experiment attached.

Thanks

experiment.zip

and best,

Kate

sebastiaan

Hi Kate @stoney ,

It sounds like Windows decided to change the application focus, which the operating system can do for different reasons, perhaps indeed including a notification. Another thing that comes to mind, because I've heard that before (though never seen it for myself), is that the experiment window loses focus if it's been inactive for a long time, for example because it is waiting for several minutes for a participant to press a button. Could that be it?

— Sebastiaan

stoney

Hi Sebastiaan,

Sorry for the late reply, we had few participants and everything ran smoothly for a while after I turned off all the Windows notifications I could find. However, it happened again today - I noticed a Windows update pending in the system tray---not sure if it was there before the experiment---but all notifications were silenced. Maybe it nonetheless had some effect. There were definitely no pop-ups though and neither I nor the participant pressed anything accidentally (which had been my other hypothesis).

Re the possibility that the participant doesn't press a button for several minutes, I don't think that could be it because it usually (?) happens just after they've responded with a key press. (It may also happen in the middle of the stimulus presentation, but I'm never concentrating when it happens so I'm not sure!).

The weird thing is, if it was just minimising the presentation window because of some notification, why would it go back to the beginning of the experiment? Anyway I have no useful details to add! I have shut off Windows updates, hopefully that helps.

Thanks,

Kate

sebastiaan

Hi @stoney ,

Sorry for the late reply, we had few participants and everything ran smoothly for a while after I turned off all the Windows notifications I could find. 

Good to hear!

The weird thing is, if it was just minimising the presentation window because of some notification, why would it go back to the beginning of the experiment?

It wouldn't! Or at least, not unless the experiment has been built such that it is able to reset itself to the beginning. But OpenSesame will not spontaneously restart an experiment. If this seemed to happen, then there was likely some other cause. For example, maybe the experiment was started twice and then when one experimental run closed, the other one was still at the beginning. Or perhaps the participant closed the experiment accidentally and then restarted it in the hope to be able to continue (and did not tell you). Could be it be something along those lines?

— Sebastiaan

stoney

Hi @sebastiaan,

Hmm I don't think so, it happens quite suddenly and for the participant to restart it would take some time. I've asked the participant each time if they accidentally pressed something other than the response keys and they've always been adamant they haven't... I feel like at least one would have owned up! Anyway the presentation paradigm shouldn't react to anything other than the response keys and Esc.

I managed to get a screenshot of it today - first the screen went totally black (almost like the power was disconnected), then it returned to what's in the screenshot: you can see there's no error message, it's just shown an experimental trial (Canvas.show()), but the presentation screen that's minimised shows the first instructions sketchpad of the whole experiment. If you zoom in enough, you can see there's no second experiment open. I double checked the screen sleep settings, but they're set to "never", so I don't think the screen went to sleep.

When you say that the experiment could possibly be built to reset itself, is there some setting that I could check for that? Or would that have to have been something I set myself?

Best,

Kate

sebastiaan

Hi @stoney ,

I see what you mean, yes. Question: I see the instructions in the little thumbnail, but can you actually reactivate the experiment and see those instructions full screen? I bet you can't right? I suspect that the window thumbnail may simply still show the instruction screen, even though the actual window has moved on. That would explain why the experiment seems to reset to the instructions, even though I don't see how this logically possible. Could that be the case?

When you say that the experiment could possibly be built to reset itself, is there some setting that I could check for that? Or would that have to have been something I set myself?

You would need to explicitly build this yourself by wrapping everything in a loop, and I can see from your screenshot that you haven't done that. So no, it seems unlikely to the point of impossibility that the experiment actually resets.

That leaves the question of why the experiment window is minimized. I honestly don't know.

— Sebastiaan

stoney

Hey @sebastiaan ,

Actually yes, you can click on the thumbnail and restart the experiment! Very weird. Anyway, thanks for all the suggestions... if I ever figure it out, I'll let you know!

Best,

Kate

stoney

Hi again @sebastiaan,

I don't know if I've solved the above problem but I have noticed one thing (posting in case it's useful to anyone): Sometimes after starting Open Sesame, I notice that the little blue wheel that replaces the mouse cursor is still spinning even after the program has been open for several minutes. At least once I remember this being the case when it returned to the start mid-experiment and might suggest something like your intuition above, that there is a second instance of Open Sesame running in the background, or that a second instance is loading, or something like this. Anyway, when I notice this happening, I close Open Sesame and restart - since then, the problem has not re-occurred (touch wood).

I'm also using Open Sesame v3 so once the current experiments finish, I'll upgrade to the v4, maybe that will fix it.

Many thanks for all the suggestions!

Best,

Kate