Doing ROI-to-ROI-FC-Analysis starting with FunImgARC-files produced warning-messages 'Warning: Attempt to write an unsupported data type to an ASCII file. Variable '...' not written to file.' when trying to write the result-files with extension '.txt' (ROICorrelation_subjxx.txt, ROICorrelation_FisherZ_subjxx.txt and ROISignals_subjxx.txt) in the result-directory 'Result/FC/FunImgARCF_ROISignals/'. Debugging showed, that the command
save([fullfile(pathstr,['ROISignals_', name]), '.txt'], 'ROISignals', '-ASCII', '-DOUBLE','-TABS') (and
in file 'y_ExtractROISignal.m' and
save([fullfile(pathstr,['ROI_', name]), '.txt'], 'SeedSeries', '-ASCII', '-DOUBLE','-TABS')
in file 'y_SCA.m'
produce the warnings and result in empty .txt-files, if the variables 'ROISignals', 'ROICorrelation' and 'SeedSeries' are single precision instead of double precision. I fixed it by the workaround
ROISignals = double(cell2mat(SeedSeries)); instead of ROISignals = cell2mat(SeedSeries); and
SeedSeries = double(cell2mat(SeedSeries)); instead of SeedSeries = cell2mat(SeedSeries);
a few lines above in the corresponding files. Maybe the best place to correct it is elsewhere. Starting analysis stages before e.g. with FunImgAR-files seems to avoid the error. The empty-.txt-files are not a severe problem, because the correspondig .mat-files are written correctly, but having the values always easy accessible outside Matlab would be nice.
save([fullfile(pathstr,['ROISignals_', name]), '.txt'], 'ROISignals', '-ASCII', '-DOUBLE','-TABS') (and
save([fullfile(pathstr,['ROICorrelation_', name]), '.txt'], 'ROICorrelation', '-ASCII', '-DOUBLE','-TABS')
save([fullfile(pathstr,['ROI_', name]), '.txt'], 'SeedSeries', '-ASCII', '-DOUBLE','-TABS')
in file 'y_SCA.m'
produce the warnings and result in empty .txt-files, if the variables 'ROISignals', 'ROICorrelation' and 'SeedSeries' are single precision instead of double precision. I fixed it by the workaround
ROISignals = double(cell2mat(SeedSeries)); instead of ROISignals = cell2mat(SeedSeries); and
SeedSeries = double(cell2mat(SeedSeries)); instead of SeedSeries = cell2mat(SeedSeries);
a few lines above in the corresponding files. Maybe the best place to correct it is elsewhere. Starting analysis stages before e.g. with FunImgAR-files seems to avoid the error. The empty-.txt-files are not a severe problem, because the correspondig .mat-files are written correctly, but having the values always easy accessible outside Matlab would be nice.
Forums
Re: DPARSF_V2.2_121225 / Warnings and empty .txt-files when ...
Thanks a lot for your reports! They are very valuable. I will fix them in the next release (probably by the end of this month).
Best,
Chao-Gan