search

klabhub / neurostim

Design and run visual neuroscience experiments using Matlab and the Psychophysics Toolbox.
MIT License
5 stars 4 forks source link

Adaptive parm fix again #192

Closed bartkrekelberg closed 2 years ago

bartkrekelberg commented 2 years ago

  1. This fixes one bug following #191 :

    Adaptive parameters assigned directly to an object were not added to the design correctly if the design contained only conditions and no factors.

  2. And it changes the way adaptive parameters are overruled by the design :

Since #191, a single condition in the design (for a plugin and parameter that also had an adaptive parameter assigned to a property directly), resulted in the adaptive parameter being ignored for all conditions. That is contrary to the way default values are handled for other (nonadaptive) parameters. I changed this now so that a design overrules an adaptive in a condition-specific manner.

See tools/adaptiveParmTest.m for a quick test/illustration.

cnuahs commented 2 years ago

Ah, Adam was chasing a bug somewhere in #191 a couple of weeks ago.

@adammorrissirrommada, is this the same issue you were dealing with?

cnuahs commented 2 years ago

tools/adaptiveParmTest.m seems to work for me, although grating.X has constant value 0 in specs(2), not 10 as per the comment. I think this is just a typo. Fixed in 99df76b.

But importantly, demos/adaptiveDemo.m seems to NOT [always] work for me in pianola mode. I think this is a problem with the model observer rather than with the adaptive plugin(s).