[solved] EyeLink

booradley

Hi folks,

I spent most of the day installing python and all the relevant plugins and libraries etc. as per the instructions on the site to accommodate EyeLink in the source code version of OpenSesame.

I can run OpenSesame from the command prompt no problem but alas when I try and add the eyelink_calibrate item the following error is presented:

Failed to load plugin 'eyelink_calibrate'.
Error: 'qteyelink_calibrate' object has no attribute 'add_checkbox_control'

We're using EyeLink 1000 if that is relevant.

Any ideas what I've done wrong?

Boo.

sebastiaan

Hi Boo,

Right, it seems that the newly updated EyeLink plug-ins use some functionality that wasn't present in OpenSesame 0.26. So the solution would be to either

• Use version 0.17 of the plug-ins
• Run the latest pre-release of OpenSesame 0.27.

Either way should work, although I think at this point you'd be better off by using the pre-release of 0.27, which should become the new stable version pretty soon.

I'd also like to point out that there is a Python-portable-based package of OpenSesame that works out of the box with the EyeLink.

Hope this helps!

Cheers,
Sebastiaan

booradley

Hi Sebastiaan,

I've made the necessary changes and I can drop the 'eyelink_calibrate' item into the experiment now without issue.

I'll test it with the EyeLink hardware later today to ensure everything is in order.

Thank you very much,

Boo.

booradley

Hi again,

I've had a crash course in how to use EyeLink today and seem to be able to successfully calibrate it through OpenSesame (although when instructed to press q to return to the menu it in fact displays a blank screen and when I press V to 'validate accuracy' -working under the assumption that the menu is accessible- it shows the fixation circle but I'm unable to initiate the process. However, as I said, calibration does appear to have worked and I can run the experiment regardless).

I've inserted 'eyelink_start_recording' at the beginning of the trial and 'eyelink_stop_recording' at the end of the trial.

I don't get any kind of error when running the experiment.

An EDF file is successfully created but when I open it there is nothing in it. No eye-tracking trajectories etc.

I've attempted to designate interest areas (working under the assumption this is what is responsible for the EDF being empty) in OpenSesame but without success. I fear I'm just not grasping how to implement interest areas.

Is it only in inline_script items where this can be done or is there an easier way to map out the coordinates of interest in eyelink_log?

An example of either would be much appreciated if at all possible.

I appreciate all the help you're providing. I plan to continue using OpenSesame in all aspects of my experimental research and in fact will likely be giving tutorials for students in the coming weeks also as I am actively promoting OpenSesame within the department.

Thanks once again,

Boo.

p.s. I will be running the behavioural experiment tomorrow morning at 11am even if I don't have the EyeLink fully functional. Eye tracking is something I can come back to after this experiment should I fail to implement it in time.

sebastiaan

I've had a crash course in how to use EyeLink today and seem to be able to successfully calibrate it through OpenSesame (although when instructed to press q to return to the menu it in fact displays a blank screen and when I press V to 'validate accuracy' -working under the assumption that the menu is accessible- it shows the fixation circle but I'm unable to initiate the process. However, as I said, calibration does appear to have worked and I can run the experiment regardless)

Right, the setup procedure is a bit flaky. Basically, the trick is to press enter when the screen stays blank, and you should eventually see the menu again. But it should work and nothing much can go wrong. Once you get a feel for it, it's quite easy to use (I think).

An EDF file is successfully created but when I open it there is nothing in it. No eye-tracking trajectories etc. I've attempted to designate interest areas (working under the assumption this is what is responsible for the EDF being empty) in OpenSesame but without success. I fear I'm just not grasping how to implement interest areas.

Interest areas are not necessary, at least not for plain recording (you might want them for analysis purposes, of course). Essentially, all you need to do is insert an eyelink_start_recording and eyelink_stop_recording item, and they should do as their names suggest. How do you know the file is empty? Is it actually a zero-sized file, or do you not see anything when you open it in a dataviewer program? I would suggest converting the EDF file to .asc with EDF2asc, which should be included or otherwise downloadable from the SR Research forum. After you have done this, you can open the resulting .asc file in a text editor and see whether it is really empty.

I appreciate all the help you're providing. I plan to continue using OpenSesame in all aspects of my experimental research and in fact will likely be giving tutorials for students in the coming weeks also as I am actively promoting OpenSesame within the department.

Good to hear!

booradley

I'm not near the eye tracking lab right now but when I opened the EDF file in Data Viewer the trials were listed but each one (when clicked) appeared 'empty'.

I say this as I have an example file from a colleague's experiment using Experiment Builder and when I click one of the trials in Data Viewer there is a playable video and mapped trajectory etc.

I have placed a few eyelink_log items before stimulus presentation with messages for the log. Could this be an issue?

Boo.

sebastiaan

I have placed a few eyelink_log items before stimulus presentation with messages for the log. Could this be an issue?

That shouldn't be a problem. The eyelink_log plug-in just adds a few messages to the log-file, but it doesn't pause recording, or anything along those lines.

I haven't encountered your problem before, but I have a hunch that it's one the following:

• The dataviewer expects the data in a slightly different format than OpenSesame records it (maybe some messages are missing?). In other words, there is data, but the dataviewer doesn't show it. I don't have any experience with the EyeLink dataviewer, so I can't say too much about it, but this seems like a possibility. As I said before, checking whether the ASCII-converted datafile is actually empty should tell you whether this is the case.

• The EyeLink is set to neither record samples (gaze position at one point in time) nor events (saccades, fixations, etc.). In the EyeLink setup screen you can select what data should be written to the EDF file, so perhaps you could check there? (It should be fairly obvious, but I can't remember the exact layout of the screen.) This would be strange though, because the plug-ins should automatically set the EyeLink so that it records both samples and events—But it never hurts to check.

Good luck!

Sebastiaan

booradley

Hi again,

There was an odd little glitch when using the keyboard attached to the tower operating OpenSesame it resulted in us not being able to record properly but we were able to calibrate.

Fast forward an hour or so and we used the keyboard attached to the EyeLink tower and it worked perfectly, validation of accuracy included. I now have some nice EDF files to play with.

I'll try the interest areas instructions again at another time but for now I'm happy that EyeLink is recording as expected.

Boo.

pkb

Hi,
We have recently shifted our lab and when we are switching the system on, the system throws an error 78 saying unable to connect the camera. Although all the cables are properly attached as per instruction manual. Kindly help.

sebastiaan

@pkb It sounds like the EyeLink PC gives this error, right? So the PC that is connected to the EyeLink camera, and not the PC that runs the experiment? If so, then this sounds like a hardware error, which could result from anything from a broken cable, to the EyeLink camera not being plugged in to the power socket, to the ethernet cables being switched around. (Or something worse, but probably not.)

So I would first recheck the set-up, and then post this question on the SR Research Support forum with a detailed description of what happens exactly. I'm sure they'll be able to help you out.