Frame rates for RSVP experiments

Support
1

Hi Chris,

I have some questions about frame rates for RSVP experiments. I don’t want to overwhelm you with a lot of questions, so if it would be better to talk verbally please let me know.

Do you know the shortest image duration we can display in a sequence of images? I don’t know how the lab’s displays are set up, but my guess is we have a 60 Hz refresh rate, giving us a lower bound of 1/60Hz ~= 17 ms images. I would also guess that image durations could be set to any multiple of the shortest duration (up to a few seconds).

Can image durations be set on a per-frame basis (i.e. variable image duration within a trial)?

I’m also interested in varying the fixation dot duration. Can this be varied in the same way as the image duration? I hope to use the standard 100 ms and decrease in increments down to 0 ms (i.e. a seamless movie-like playback).

How would the image and fixation durations be encoded in the experiment?

In addition to RSVP, I’m hoping to display movies. I suppose I could cut a movie into individual frames and present them just like RSVP. Alternately, is there a way to play back intact movies (i.e. not cut into frames) and have the same timing accuracy/precision as RSVP? I’m also interested in varying the frame rate/image duration in the same way as RSVP. My guess is the best option is to cut the movie into frames, but I wanted to get an idea of the options.

Thanks in advance.

2

Hi Rob,

Do you know the shortest image duration we can display in a sequence of images? I don’t know how the lab’s displays are set up, but my guess is we have a 60 Hz refresh rate, giving us a lower bound of 1/60Hz ~= 17 ms images.

You’re correct that the shortest duration MWorks can display an image (or any other visual stimulus) is one refresh period. The length of the refresh period depends on the specific display you’re using. I don’t manage the hardware in the DiCarlo lab’s setups, but I wouldn’t assume that all the displays run at 60Hz.

I would also guess that image durations could be set to any multiple of the shortest duration (up to a few seconds).

That’s correct, but if you really want the durations to be accurate to a specific frame count, you’ll probably need to use a frame list stimulus. You might also be able to do it with render actions.

If you can describe your protocol in more detail, I can make an example experiment that demonstrates one or both of these techniques.

Can image durations be set on a per-frame basis (i.e. variable image duration within a trial)?

It should be possible, using one of the techniques I mentioned above.

I’m also interested in varying the fixation dot duration. Can this be varied in the same way as the image duration?

Yes, you should be able to use the same technique for any visual stimulus.

How would the image and fixation durations be encoded in the experiment?

That depends on the details of your experiment. Again, if you can describe the protocol in detail, I can provide example code.

In addition to RSVP, I’m hoping to display movies. I suppose I could cut a movie into individual frames and present them just like RSVP. Alternately, is there a way to play back intact movies (i.e. not cut into frames) and have the same timing accuracy/precision as RSVP? I’m also interested in varying the frame rate/image duration in the same way as RSVP. My guess is the best option is to cut the movie into frames, but I wanted to get an idea of the options.

MWorks does support true video playback, but it doesn’t let you vary the playback rate. To do what you want, I think you would need to cut the movies into one image per frame.

Cheers,
Chris