search

cvnlab / GLMsingle

A toolbox for accurate single-trial estimates in fMRI time-series data
BSD 3-Clause "New" or "Revised" License
101 stars 45 forks source link

feature request: provide mask for top 1% voxels related to a runwiseFIR.png #121

Closed andreifoldes closed 11 months ago

andreifoldes commented 11 months ago

Hello,

Much thanks for the cool diagnostics plot! I think it would help with following up on suspicious runs if users had access to the binary mask(s) of the voxels going into the runwiseFIR plot.

andreifoldes commented 11 months ago

Upon further reflection, these could be created by thresholding the R2 map, although that's aggregated across runs.. still it would largely overlap right?

kendrickkay commented 11 months ago
image

Yes, that is a good idea and is on my list of things to implement

kendrickkay commented 11 months ago

This has been implemented in the matlab version. (Python to soon follow). See:

https://github.com/cvnlab/GLMsingle/commit/31f9eaa46e6f8c01cde49315826a0aef5cc674f1