[FieldTrip] Clarification on plotting WPLI

[Cristiano Micheli]

Dear Yusuf and Peter,

My experience is also that negative squared wPLI values should not be
considered, as they constitute a side-effect of the debiasing technique. I
would rather clip them to zero, or, better, assign them a NaN (since the
result is undetermined).

I hope this helped.
Cris



On Fri, Feb 19, 2016 at 3:04 PM, Prasandhya Astagiri Yusuf <
prasandhya.a.yusuf at gmail.com> wrote:

> Hi Peter,
> sorry for the (very) late response, I guess you have already figured the
> answer by now. Anyhow, in case someone has similar problem, I try to share
> my experience here in the forum.
>
> I recently computed the debiased WPLI and experienced the same problem. I
> also found this negative value of debiased WPLI quite frequently asked in
> this mailing list, even Mr. Martin Vinck himself has already responded to
> it on June 2011.
> So as stated in the paper (Vinck et al. 2011), PLI and WPLI result has the
> range from 0 to 1. But the debiased WPLI could yields negative values as a
> consequence of debiasing method. One of his latest paper (Phillips, 2014),
> they  mentioned in the Materials and Methods that "the debiased WPLI
> estimator (Vinck et al. 2011), ranging from zero (*negative values can
> incidentally occur because of limited sampling*) to one (maximum
> coherence)".
> So in my opinion, we can neglect this negative values and convert them to
> zeros. But surely, we should not take the absolute value from it.
>
> Best,
> Sandhy
>
>
> Prasandhya A. Yusuf
>
> *Medizinische Hochschule Hannover*
>
> Experimentelle Otologie / OE 8891 /
>
> VIANNA / M13
>
> Feodor-Lynen-Str. 35
>
> D-30625 Hannover
>
>
>
> On Mon, Dec 21, 2015 at 8:39 PM, Peter Lyons <plyons at udel.edu> wrote:
>
>> Hello Everyone,
>>
>> I have a question regarding the interpretation of negative values in the
>> debiased wPLI. Although the reference page on the debiased wPLI notes that
>> estimates are computed by squaring the wPLI values, I and others have
>> noticed that some of the values in the debiased wPLI spectrum are negative.
>> How should these negative values be interpreted in reference to 0?
>>
>> To view the results of the debiased wPLI connectivity analysis
>> (ft_connectivity analysis with cfg.method = 'wpli_debiased'), I've been
>> using the ft_topoplotER function. However, I'm concerned that the default
>> scaling (zlim = 'maxmin') of the plot does not seem to register the
>> negative values in the debiased wPLI spectrum. That is, the negative values
>> are not being plotted. It is curious, however, that the ft_topoplotER
>> does register the negative values in the *non-debiased *wPLI spectrum.
>>
>> Is there a reason why the ft_topoplotER function is treating the negative
>> values from the debiased vs. non-debiased wPLI spectrums differently? Is
>> there perhaps a more appropriate plotting function I should be using to
>> view the debiased wPLI output?
>>
>> I would greatly appreciate any guidance in this.
>>
>> Thank you!
>>
>> Peter
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20160219/5ca8feb2/attachment-0002.html>