[FieldTrip] coherence estimation with "FT_FREQANALYSIS_MTMWELCH"

[David Meng]

Dear fieldtrippers,


I just found a description page of the function "ft_freqanalysis_mtmwelch"(http://www.fieldtriptoolbox.org/reference/ft_freqanalysis_mtmwelch).


Seems this function has been taken out of the package but as it is described on the webpage above, this Welch's overlapped averaged periodogram method can be implemented as:


[freq] = ft_freqanalysis(cfg, data)


by specifying 'cfg.method = mtmconvol' and use the options of this function to specify   the length of the time windows, the amount of overlap, and the amount   of spectral smoothing (in case of dpss tapers) per window.


I have tried to implement this method as suggested, however in the field of the output "freq", the power spectrum and cross spectrum has a dimension as 'chan_freq_time' which means it didn't calculate the average value over all time windows.


Would anyone suggest a way to enable the averaging? I have tried to average them manually but the topoplot from following 'ft_connectivityanalysis' could not be shown corrected.


Best regards

David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20160225/3518ce40/attachment-0001.html>