search

Mouse-Imaging-Centre / RMINC

Statistics for MINC volumes: A library to integrate voxel-based statistics for MINC volumes into the R environment. Supports getting and writing of MINC volumes, running voxel-wise linear models, correlations, etc.; correcting for multiple comparisons using the False Discovery Rate, and more. With contributions from Jason Lerch, Chris Hammill, Jim Nikelski and Matthijs van Eede. Some additional information can be found here:
https://mouse-imaging-centre.github.io/RMINC
Other
22 stars 17 forks source link

vertexTFCE, allow one weight file per subject? #144

Closed gdevenyi closed 7 years ago

gdevenyi commented 7 years ago

This I think is just a further generalization of the option allowing surface to define the local weights.

Rather than having RMINC calculate the areas using the surface, I'd like to provide the local surface area measures (as calculated by the Voronoi method, provided by CIVET).

In surfaces with more uneven triangulation the Voronoi method provides a better surface area estimate than the simple area.

cfhammill commented 7 years ago

Hi Gabe, I think for a singular voronoi tessellation it's pretty to trivial to implement on the user side: Just do weights = as.numeric(readLines("voronoi.txt"))

I had considered the multi-subject case but thought that it didn't really make any sense. Individuals don't contribute directly to the TFCE stat, only through the unifying statistic map. So one tessellation per stats map maybe, but not per subject. Does that make sense, or can you think of a use case where each individual would get a different weighting (maybe some exercise in what would the enhancement look like had it been run on each individual surface?)

gdevenyi commented 7 years ago

Maybe I'm misinterpreting surface? 

For the matrix and vertexLm cases, either a single surface object may be passed and used for each individual, or a vector of file names

Does that say I can provide one surface per subject? Or am I misinterpreting that sentence?

cfhammill commented 7 years ago

I guess I did add that ability, but again I don't think it makes any sense to use it. For the vertexLm case it would be one surface per statistic (not subject), and for the matrix case it would be one per column (whatever the columns are).

I guess for now I can add passing of one weight per surface until I a good argument to remove the feature comes up.

gdevenyi commented 7 years ago

Okay, that makes more sense. I didn't see why one surface per subject made sense since they "should" all correspond for comparisons to make sense.