[FieldTrip] ft_connectivityanalysis

Schoffelen, J.M. (Jan Mathijs) janmathijs.schoffelen at donders.ru.nl
Thu Mar 3 19:50:11 CET 2022


Hi Sofia,

Your issue is most likely caused by the fact that the input data that you use for ft_connectivityanalysis consists of just a single trial worth of data. In general, the computation of a correlation coefficient needs multiple paired observations. If there’s just a single data duplet, then the correlation will by definition be 1. FieldTrip’s connectivity routines require the observations to be represented as trials, i.e. a powspctrm that has a ‘dimord’ of ‘rpt_chan_freq(_time)’

best wishes,
Jan-Mathijs


On 3 Mar 2022, at 08:59, Stephen Whitmarsh via fieldtrip <fieldtrip at science.ru.nl<mailto:fieldtrip at science.ru.nl>> wrote:

Dear Sofia,

It might be that something is either off with the input data, or your configuration of ft_connectivityanalysis.

A) I would first check that the two channels of data you are using are not (highly) correlated, e.g. by just plotting (part of) them together.
B) Due to the fact you are using intracranial data of very closely spaced electrodes, more detail on their preprocessing would also be helpful; e.g. are you using bipolar references/CSD or a common references to both?

Getting 1's indicate something is more probably off with your analyses though. To be able for people to evaluate that:
C) Share your configuration to both ft_connectivity, and the previous ft_freqanalysis, so that people can evaluate whether you are calling those functions correctly. Preprocessing steps will also be helpful.
D) Do you get any error/warning messages at any point?

I hope this helps,
Stephen

On Wed, 2 Mar 2022 at 12:59, CARRION-FALGARONA Sofia via fieldtrip <fieldtrip at science.ru.nl<mailto:fieldtrip at science.ru.nl>> wrote:
Dear Community,

I am Sofia and I work at the Brain Paris institute. I am currently looking at rodent’s brain activity within different brain’s region simultaneously by using deep multi-electrodes.
I am trying to use ft_correlationanalysis function in order to compare the power correlation of different channel through the time.
I was expecting to observe some fluctuation in the correlation but I obtained a matrix of  1…
Can someone help me in order to get the correct correlation between 2 channels ?



stat  = ft_connectivityanalysis(cfgtemp, timefreq_alldata);
disp( stat);

>> label: {2×1 cell}
   dimord: 'chan_chan_freq_time'
   powcorrspctrm: [2×2×50×801 double]
   time: [1×801 double]
   freq: [1×50 double]
   dof: 1
   cfg: [1×1 struct]

data_corr_temp  = stat.powcorrspctrm(1,1,:,:);
data_corr       = permute(data_corr_temp ,[3 4 1 2 ]);
disp(data_corr);

>> 1 1 1 1 1 1 1 1 1 1 . . . => actually return a Matrix of 50x801 of 1

I thank you for your help !

Best Regards,

Sofia
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202<https://urldefense.com/v3/__https://doi.org/10.1371/journal.pcbi.1002202__;!!HJOPV4FYYWzcc1jazlU!5cfJ4Ncc5pfsMd3XfB7rN4IgV4Rihfl6_cc-r-4sNFN0L0u3yrapN_7AC7eqD-mC0dAWbCLglwvGsDh7zXHZ6FPhOOXwzeD2H7oSFg$>
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://urldefense.com/v3/__https://doi.org/10.1371/journal.pcbi.1002202__;!!HJOPV4FYYWzcc1jazlU!5cfJ4Ncc5pfsMd3XfB7rN4IgV4Rihfl6_cc-r-4sNFN0L0u3yrapN_7AC7eqD-mC0dAWbCLglwvGsDh7zXHZ6FPhOOXwzeD2H7oSFg$

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220303/b2618401/attachment.htm>


More information about the fieldtrip mailing list