[Homer-users] svd encountering NaN

[Andrea Vitale]

Dear all,

I'm encountering some problem with Homer2 following the preprocessing
stream presented in the webseminar of April:

- when performing the enPCAfilter function, Homer2 reports an error because
of NaN values.
Going step by step through the preprocessing stream we found that the NaN
values are generated as the output of the hmrMotionSpline function:
Here enclosed a figure (1) representing the dodSpline as timepoints in
function of the channels (nChan = 40): the two vertical red bars indicates
that there're some not numerical values for channels 4 and 24.
The second figure (2) shows dodSpline values across the time (each channel
is represented in a different colour).

To avoid the NaN values shall we replace them with a zero or with the last
valid number? What are the possible consequences in each of these cases?

Why do you think the NaN values are created? Is it connected with the data
collection and thus with the input data?

FYI: the data I'm using were collected on newborns with a ISS Imagent
(source-detector distance = 18mm).

Looking forward to your suggestions,
best


andrea vitale
master student - CIMeC
University of Trento
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/homer-users/attachments/20140925/cbbe6f23/attachment.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fig1.PNG
Type: image/png
Size: 60390 bytes
Desc: not available
Url : http://mail.nmr.mgh.harvard.edu/pipermail/homer-users/attachments/20140925/cbbe6f23/attachment.png 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fig2.PNG
Type: image/png
Size: 72384 bytes
Desc: not available
Url : http://mail.nmr.mgh.harvard.edu/pipermail/homer-users/attachments/20140925/cbbe6f23/attachment-0001.png