[Eeglablist] Error message during time-frequency analysis

[Scott Makeig]

Amy -  The ITC phase coherence functions divide by the amplitude values, so
/0 blows up. A more general approach might be to add a small bit of white
noise to the data -- and investigate why the zeroes occur...

Scott Makeig

On Thu, May 26, 2011 at 11:23 AM, Amy Arlene Nash <Amy.Nash at colorado.edu>wrote:

>  Dear Arno -
>
> Thanks so much for the time in responding.  However, I am already using
> epoched files, not continuous data (.eeg is always epoched in Neuroscan – if
> it were continuous, it would be a .cnt file).  The problem seems to center
> upon the script’s inability to deal with zeroes in the data.  The only
> workaround I’ve found so far (since writing earlier this month) is to add a
> very small amount to the data to eliminate any possibility of zeroes (so
> EEG.data=EEG.data+.00001 for example), and then it appears to analyze and
> plot well.  Not ideal, but it is better than what I was dealing with before.
>
> With thanks,
> Amy
>
>
>
> On 5/25/11 11:23 PM, "Arnaud Delorme" <arno at ucsd.edu> wrote:
>
> Dear Amy,
>
> I think the problem arises because you are trying to compute ERSP on
> continuous data. You need to extract data epochs first.
> Best,
>
> Arno
>
> On May 5, 2011, at 2:34 PM, Amy Arlene Nash wrote:
>
> Thank you in advance for any help you all can offer with this question.
>
> I have been running time-frequency analyses on a number of datasets which I
> import from Neuroscan .eeg files.  Importing is no problem and most of the
> time, I am able to get the analysis to run without issue, and lovely plots
> are the result.  However, some of the time I get a warning about dividing by
> zero, followed by a warning that the imaginary parts of complex X and/or Y
> arguments are ignored.
>
> Ultimately, the ERSP usually plots well, but the ITC is all dark blue, with
> an error message that says: Bad property value found, Object name: axes,
> Property name: ‘YLim’, Values must be increasing and non-NAN.
>
> Is there something I could change about my imported data files or analysis
> parameters that would allow me to avoid this error?  These EEG recordings
> were all acquired in the same way, and have similar properties overall... I
> haven’t been able to figure out why some files will analyze and plot and
> others won’t.  I apologize for my lack of general MATLAB troubleshooting
> knowledge, and again, appreciate your input.  I would hate to lose some of
> my sample to analysis errors like this.
>
> Many thanks,
> Amy Nash
>
> Amy Nash, M.S., CCC-A
> Doctoral Candidate, University of Colorado at Boulder
> Brain and Behavior Laboratory
> 503-926-4919 (cell)
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>
>
>
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>



-- 
Scott Makeig, Research Scientist and Director, Swartz Center for
Computational Neuroscience, Institute for Neural Computation & Adj. Prof. of
Neurosciences, University of California San Diego, La Jolla CA 92093-0559,
http://sccn.ucsd.edu/~scott
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20110605/883279f1/attachment.html>