Hi
can you please email the report.log file. It seems that the SVD fails.
This can happen for mainly two reasons- i) the image dimensions are
wrong (did you use avwmerge -t ?) or (ii) the image intensities are too
similar (e.g. because the data has been converted to int). You can
check these things using avwsize, avwstats and/or by loading the data
into fslview. If you put the data somewhere for me to d/l I'll have a
quick llook
cheers
c
On 13 Jul 2004, at 03:03, Dhananjay Thiruvady wrote:
> Christian,
>
> This is going back a few weeks ago. I am attempting to identify the
> independent components in a group of images. This is the error I get
> after using avwmerge:
>
> Starting PCA ...
>
> An exception has been thrown
> Runtime error:- detected by Newmat: process fails to converge
>
> MatrixType = Diag # Rows = 2; # Cols = 2; lower BW = 0; upper BW = 0
> Trace: Evalue(tql2).
>
> Does this imply that the independent components cannot be identified?
>
> Regards,
>
> DJ
>
> Christian Beckmann <[log in to unmask]> wrote:
>> DJ
>>
>> time series analysis requires time series - all our functional tools
>> work
>> on 4D files. It seems that you've been trying to run it on a single 3D
>> volume:
>>
>>> Melodic Version 2.0
>>>
>>> Melodic results will be in /home/norm_volumes_1/nepi_vol_0002.ica
>>>
>>> Reading data file filtered_func_data ... done
>>> Create mask ... done
>>> Excluding voxels with constant value
>>> Data size : 1 x 510340
>>
>> You will need to create a 4d file first (avwmerge -t does this). If
>> the
>> file should have been 4D already, check the header information
>> (avwhd) to
>> see why it appears as a 3d volume
>>
>> ta
>> c
>>
--
Christian F. Beckmann
Oxford University Centre for Functional
Magnetic Resonance Imaging of the Brain,
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
Email: [log in to unmask] - http://www.fmrib.ox.ac.uk/~beckmann/
Phone: +44(0)1865 222782 Fax: +44(0)1865 222717
|