This is actually quite simple, just create the following design:
where the Gaussian dipoles are defined as follows:
and the Glass patterns organization is specified using a MEF stimulus with the new "Glass Patterns" distribution option which makes sure that the dipoles orientation is tangent to concentric circles:
Thank you - this was very helpful. To go a step further with this, it seems that I can vary the % signal by having a percentage of the dipoles with a glass distribution and the remaining percentage of dipoles as a random distribution. Is there a way to simultaneously vary these two percentages using a staircase procedure or method of constant stimuli to obtain a threshold?
If you define a variable 'coherence' as the dependent parameter of your method, then simply specify '[coherence]' for the % number of Glass dipoles and '100-[coherence]' for the % number of random dipoles!
I used the above instructions to make a glass pattern stimuli and slotted it into a 2AFC procedure using a Bayesian method. Coherence (orientation jitter) is the judged parameter.
It works perfectly when using a temporal presentation but I need a spatial presentation. When set to this option the two glass patterns are laid over one another in the centre of the screen.
When I change the stimuli to something else (e.g. RDK Gaussian blob) the spatial presentation works a treat. Therefore I am sure it is something to do with the glass pattern stimuli. Is there something within the design that is overriding the spatial presentation and forcing the glass pattern stimuli to be presented over laid?
Do you know how I could achieve the spatial presentation using the glass pattern stimuli?
Is it possible to get a patch or an update for the bug fix? I need to be able to run the described experiment in the next few weeks for a masters project.
I'm having further problems with the Glass pattern experiment and am unsure if it's another bug or not. The last fix worked great.
I am now replicating the experiment first described but with the judged parameter now as contrast, not orientation jitter (2AFC procedure using a Bayesian method).
When the presentation type is temporal the experiment works fine - each glass pattern has a different contrast level. However, when the presentation type is spatial both glass patterns have the same contrast level meaning discrimination is impossible. It varies from trial to trial but both glass patterns always have the same contrast level.
I tried generating glass pattern with the above instruction. I have adopted a staircase method with nAFC approach, so that I can measure a threshold. I wanted to have a presentation of two side by side stimuli in a random order(one with glass pattern and the other one without glass pattern) with vertical split. But unfortunately I ended up in having two similar glass patterns with a vertical split. Could you please help me in getting which I actually want. I am not sure whether to make any changes in the coding of the "custom" section.
It would be great if you could provide me a sequential instructions with a couple of screenshots.