search

ANTsX / ANTsR

R interface to the ANTs biomedical image processing library
https://antsx.github.io/ANTsR
Apache License 2.0
127 stars 35 forks source link

Question about SiMLR usage #329

Closed ntustison closed 3 years ago

ntustison commented 3 years ago

Hey @stnava ,

I'm doing an analysis with three modalities (cortical thickness, fALFF, and FA) mapping them all to a common template and using the same template mask for all three (as I believed we've done in the past). I had some initial issues but was able to trace it to the scaling for the cortical thickness view where NaNs were being produced outside the cortex (but within the mask). Am I correct in remembering that we used the same mask for all modalities and, if so, how does one handle those voxels which have 0 thickness since they're outside the cortex but also within the template mask? Thanks.

stnava commented 3 years ago

for thickness, just use a mask that excludes nans

stnava commented 3 years ago

or change the default scaling options - see ?simlr. scale options

ntustison commented 3 years ago

Got it. Thanks.

ntustison commented 3 years ago

Looks like it worked, @stnava . Haven't explored it thoroughly yet but was able to do some statistical testing. This is awesome. 

initializing with random matrix:  2 columns
[1] "     <0> BUILD-V <0> BUILD-V <0> BUILD-V <0> BUILD-V <0>    "
[1] "initialDataTerm: 1.88198916901763  <o> mixer: ica  <o> E:  regression"
[1] 0.06741957 0.06746118 0.06748217
[1] "Iteration: 1 bestEv: 0.273646518738461 bestIt: 1 CE: 0.273646518738461"
[1] 0.01764212 0.01809735 0.01824297
[1] "Iteration: 2 bestEv: 0.267428370033175 bestIt: 2 CE: 0.267428370033175"
[1] 0.05790177 0.03047702 0.03110222
[1] "Iteration: 3 bestEv: 0.263214563913766 bestIt: 3 CE: 0.263214563913766"
[1] 0.1701345 0.1112153 0.1127311
[1] "Iteration: 4 bestEv: 0.248118898090993 bestIt: 4 CE: 0.248118898090993"
[1] 0.3559424 0.3939283 0.4659427
[1] "Iteration: 5 bestEv: 0.184365193228254 bestIt: 5 CE: 0.184365193228254"
[1]  0.4930612  0.1299358 -0.7229219
[1] "Iteration: 6 bestEv: 0.092517667408483 bestIt: 6 CE: 0.092517667408483"
[1]  0.5039931 -0.3495042  0.7845724
[1] "Iteration: 7 bestEv: 0.0389201582643552 bestIt: 7 CE: 0.0389201582643552"
[1]  0.50239191474 -0.00006594644  0.33936210381
[1] "Iteration: 8 bestEv: 0.0081801405311484 bestIt: 8 CE: 0.0081801405311484"
[1]  0.5008925477 -0.0009299671 -0.0172962660
[1] "Iteration: 9 bestEv: 0.00233603246774376 bestIt: 9 CE: 0.00233603246774376"
[1] 0.500741749 0.002541721 0.030242872
[1] "Iteration: 10 bestEv: 0.00221275259030574 bestIt: 10 CE: 0.00221275259030574"