[VoxBo] Question on setting up a GLM with a long condition function

[Daniel Kimberg]

Lila wrote:
> I was wondering, is there any way to circumvent this problem? Do I
> have any alternative ways of going about this?

It's hard to suggest alternatives without any information about what you're trying to do.  But assuming this is the best or only way to model your data, I can think of one workaround that might help.  The GLM designer has an "upsampling rate," which actually means the sampling rate of the upsampled vectors used internally.  By default it's set to 100, but if you don't need that finer temporal resolution to model your data appropriately, you can set it higher, as high as your TR, and it will reduce the length of those vectors by an order of magnitude or so.

dan