Hi,
I believe your 3 mystery rows are x y z coordinates, not sure about
the mystery column of zeros.
Dan
On 7/21/07, Daniel Irlam <[log in to unmask]> wrote:
> Hello FSL users
>
> I am a new to the FSL community, so please bear with me if I am asking
> questions that are already out there.
> But after reading the FSL web pages and searching the FSL mailing lists I
> have some questions about avwmeants:
>
> (1) If I have a 4d epi functional time series, thus:
>
> $ avwinfo filtered_func_data.nii.gz
> data_type FLOAT32
> dim1 64
> dim2 64
> dim3 43
> dim4 193
> datatype 16
> pixdim1 3.0000000000
> pixdim2 3.0000000000
> pixdim3 3.0000000000
> pixdim4 3.0000000000
>
> i.e. 193 volumes in the time series
> and a matched mask, thus:
>
> $ avwinfo mask.nii.gz
> data_type FLOAT32
> dim1 64
> dim2 64
> dim3 43
> dim4 1
> datatype 16
> pixdim1 3.0000000000
> pixdim2 3.0000000000
> pixdim3 3.0000000000
> pixdim4 1.0000000000
>
> $ avwstats++ mask.nii.gz -V
> 447 12069.000000
>
> i.e. 447 voxels in the mask
>
> and I use avwmeants with the --showall option:
> $ avwmeants -i filtered_func_data.nii.gz -m mask.nii.gz --showall -v -o
> tmp.txt
> Number of voxels used = 447
>
> then I was expecting the resulting output file tmp.txt to have 193 rows (for
> the number of volumes) and 447 columns (for the number of voxels in the
> mask). But instead it has 196 rows and 448 columns. It seems like the first
> 3 rows and the last column of every row are 'extra' data, with the last
> column always being zero. Is this correct? If so, what is this extra data?
>
>
> (2) I can see how to extract a mean time series using data only within the
> mask:
>
> $ avwmeants -i filtered_func_data.nii.gz -m mask.nii.gz -o average_ts.txt
>
> But can I reverse this process using any of the avwutils tools? If, say, I
> now want to import timeseries data that I have in a text file (like
> average_ts.txt above) into a 4d data set at a particular voxel location, is
> it possible to do this? (Supposing for example, I wanted to create a
> synthetic data set with my own timeseries so that I could test out various
> modelling stategies). Or is this outside the scope of the avwutils tools and
> I should write my own code in matlab etc?
>
>
> many thanks
>
> Dan
>
>
> --
> Daniel Irlam Ph.D
>
|