help > Multiple VOIs within a single gPPI
Showing 1-3 of 3 posts
Display:
Results per page:
Nov 30, 2017  11:11 PM | Michelle Eisenberg - Stanford University
Multiple VOIs within a single gPPI
I would like to include multiple VOIs within a single gPPI model. When I tried putting multiple regions in for P.VOI and P.region, I got the following errors: 

ERROR 1: Program will exit. VOI is not specified.
ERROR 2: Program will exit. Region name must be a string.
One or more inputs are not correct.

I am not sure whether I am specifying the multiple regions incorrectly or whether this is just not doable within gPPI, and instead, it is necessary to run a separate gPPI model for each VOI of interest.

Here is how I have tried doing this:
regionfile={'/Users/michelleeisenberg/Documents/Postdoc/EtkinLab/Cohen/Analysis/Task_Analysis/MSIT/Test8/r_s1_sphere_2mm.nii' '/Users/michelleeisenberg/Documents/Postdoc/EtkinLab/Cohen/Analysis/Task_Analysis/MSIT/Test8/r_s3_sphere_2mm.nii'}

region={'s1_gPPI' 's3_gPPI'}

P.subject='2310088'
P.directory='/Users/michelleeisenberg/Documents/Postdoc/EtkinLab/Cohen/Analysis/Task_Analysis/MSIT/Test8'
P.VOI=[regionfile]
P.Region=region
P.Estimate=1
P.contrast=0
P.extract='eig'
P.Tasks={'1' 'Con' 'InCon'}
P.Weights=[]
P.analysis='psy'
P.method='cond'
P.CompContrasts=1
P.Weighted=0
P.Contrasts(1).left={'InCon'}
P.Contrasts(1).right={'Con'}
P.Contrasts(1).STAT='T'
P.Contrasts(1).Weighted=0
P.Contrasts(1).MinEvents=1
P.Contrasts(1).name='inc-con'
P.outdir='/Users/michelleeisenberg/Documents/Postdoc/EtkinLab/Cohen/Analysis/Task_Analysis/MSIT/Test8'
PPPI(P)

Thank you in advance for your help.
Dec 16, 2017  04:12 PM | Donald McLaren
RE: Multiple VOIs within a single gPPI
P.VOI can only contain a single VOI. For a second VOI, use P.VOI2.

P.region should always contain a single string. The program should parse the two region names.

Before trying this, consider the analysis that you are trying to perform.

By including 2 VOI, you are conducting a Pyscho-physio-physiological interaction analysis or physio-physiological analysis. If you want a psychophysiological interaction, you need to loop over the P.VOI an P.region fields and estimate it separately for each model. There is a speed improvement if you can set P.WB=1 as it will store the data in a mat-file for faster reading on the second VOI.

Best,
Donald
Dec 17, 2017  01:12 AM | Michelle Eisenberg - Stanford University
RE: Multiple VOIs within a single gPPI
Thank you. That clears it up for me.

Michelle