help
help > RE: Discrepancy between GUI and Batch processing
May 13, 2016 04:05 AM | Alfonso Nieto-Castanon - Boston University
RE: Discrepancy between GUI and Batch processing
Hi Vikram,
Yes, both gui and batch processing use the exact same functions so you should be getting the same results. Could you please send me both of your conn_*.mat files so I can take a quick look to see if I can identify any potential differences between those projects?
Thanks
Alfonso
Originally posted by Vikram Rao:
Yes, both gui and batch processing use the exact same functions so you should be getting the same results. Could you please send me both of your conn_*.mat files so I can take a quick look to see if I can identify any potential differences between those projects?
Thanks
Alfonso
Originally posted by Vikram Rao:
Hi Conn Users,
I have been wanting to batch the gui setup using conn_batch. I m using conn 13 with spm8 and both gui and batch run without any errors, however, the results dont match. The only two things i m changing from defaults in both gui and batch is the following: Changing units from percent signal change to 'raw' and i am doing only roi-roi analysis. I am not sure why the final correlation matrices are different. I am looking within the first level folder. Please advise.
Following is my code:
%% CONN batch using CONN_13
clc
clear all
close all
[dirlist]=textread('/Volumes/analysis/CONN/CONN_auto/data/file_list.txt','%s');% text file that contains the path to each subject's folders
addpath(genpath('/scripts/matlab_shared_toolboxes/conn/')) % CONN_13
spm8
%% CONN setup
batch.filename='CONN_auto.mat';
batch.Setup.isnew=1;
batch.Setup.overwrite='Yes';
batch.Setup.acquisitiontype=1;
batch.Setup.voxelmask=1;
batch.Setup.voxelresolution=1;
batch.Setup.nsubjects=9;
batch.Setup.RT=2;
batch.Setup.analyses=[1]; %Only ROI-to-ROI pipelines (changed from the default)
batch.Setup.analysisunits=2; % raw unit (changed from default: percent signal change)
batch.Setup.voxelresolution=1;
batch.Setup.roiextract=2;
batch.Setup.overwrite='Yes';
for isub=1:length(dirlist) %9
rootdir=dirlist{isub};
batch.Setup.functionals{isub}{1}= spm_select('ExtFPList',fullfile(rootdir,'Resting'),'^rest_dn_w.*s\.nii$',1:216);
batch.Setup.structurals{isub}=spm_select('ExtFPList',fullfile(rootdir,'anat'),'^wFSP.*R\.nii$');
batch.Setup.covariates.files{1}{isub}{1}=spm_select('FPList',fullfile(rootdir,'Resting'),'^art_regression_outliers_and_movement.*\.mat$');
batch.Setup.covariates.names={'motionart'};
batch.Setup.conditions.names={'rest'};
batch.Setup.conditions.onsets{1}{isub}{1}=0; batch.Setup.conditions.durations{1}{isub}{1}=inf; % rest condition (all sessions)
end
batch.Setup.done=1;
%% CONN preprocessing (all defaults)
batch.Preprocessing.filter=[0.0080, 0.0900];
batch.Preprocessing.detrending=1;
batch.Preprocessing.despiking=1;
batch.Preprocessing.overwrite='Yes';
batch.Preprocessing.done=1;
%% CONN Analysis (all defaults)
batch.Analysis.done=1;
batch.Analysis.overwrite='Yes';
%% Run the batch
conn_batch(batch)
I have been wanting to batch the gui setup using conn_batch. I m using conn 13 with spm8 and both gui and batch run without any errors, however, the results dont match. The only two things i m changing from defaults in both gui and batch is the following: Changing units from percent signal change to 'raw' and i am doing only roi-roi analysis. I am not sure why the final correlation matrices are different. I am looking within the first level folder. Please advise.
Following is my code:
%% CONN batch using CONN_13
clc
clear all
close all
[dirlist]=textread('/Volumes/analysis/CONN/CONN_auto/data/file_list.txt','%s');% text file that contains the path to each subject's folders
addpath(genpath('/scripts/matlab_shared_toolboxes/conn/')) % CONN_13
spm8
%% CONN setup
batch.filename='CONN_auto.mat';
batch.Setup.isnew=1;
batch.Setup.overwrite='Yes';
batch.Setup.acquisitiontype=1;
batch.Setup.voxelmask=1;
batch.Setup.voxelresolution=1;
batch.Setup.nsubjects=9;
batch.Setup.RT=2;
batch.Setup.analyses=[1]; %Only ROI-to-ROI pipelines (changed from the default)
batch.Setup.analysisunits=2; % raw unit (changed from default: percent signal change)
batch.Setup.voxelresolution=1;
batch.Setup.roiextract=2;
batch.Setup.overwrite='Yes';
for isub=1:length(dirlist) %9
rootdir=dirlist{isub};
batch.Setup.functionals{isub}{1}= spm_select('ExtFPList',fullfile(rootdir,'Resting'),'^rest_dn_w.*s\.nii$',1:216);
batch.Setup.structurals{isub}=spm_select('ExtFPList',fullfile(rootdir,'anat'),'^wFSP.*R\.nii$');
batch.Setup.covariates.files{1}{isub}{1}=spm_select('FPList',fullfile(rootdir,'Resting'),'^art_regression_outliers_and_movement.*\.mat$');
batch.Setup.covariates.names={'motionart'};
batch.Setup.conditions.names={'rest'};
batch.Setup.conditions.onsets{1}{isub}{1}=0; batch.Setup.conditions.durations{1}{isub}{1}=inf; % rest condition (all sessions)
end
batch.Setup.done=1;
%% CONN preprocessing (all defaults)
batch.Preprocessing.filter=[0.0080, 0.0900];
batch.Preprocessing.detrending=1;
batch.Preprocessing.despiking=1;
batch.Preprocessing.overwrite='Yes';
batch.Preprocessing.done=1;
%% CONN Analysis (all defaults)
batch.Analysis.done=1;
batch.Analysis.overwrite='Yes';
%% Run the batch
conn_batch(batch)
Threaded View
| Title | Author | Date |
|---|---|---|
| Vikram Rao | May 9, 2016 | |
| Alfonso Nieto-Castanon | May 13, 2016 | |
