[FieldTrip] TFR and DICS in Gamma Band

[Philip Cho]

Hello FieldTrip community,

I'm interested in looking at how the power of frequencies in 30-90Hz
evolves over time after an event, and eventually I would like to perform
source reconstruction to see how this power changes in certain ROIs. I have
an idea of how to approach this, but I am slightly confused on capturing
this range of frequencies. One discussion post (
https://mailman.science.ru.nl/pipermail/fieldtrip/2015-April/035024.html)
suggests that I specify one frequency at the center of the band (
cfg.foilim = [60,60] ) and use frequency smoothing over the whole
range ( cfg.tapsmofrq
= 30 ) to analyze the 30-90Hz band. However, a post from another discussion
(
https://mailman.science.ru.nl/pipermail/fieldtrip/2015-September/009589.html)
seems
to analyze multiple frequencies ( cfg.foi = 30:2:90 ) in conjunction with
frequency smoothing. My assumption is that if I use the latter, I would
average the power across the frequencies to analyze the overall band?

I'm curious which approach is best. I tried the former approach, as the
latter approach seemed more computationally intensive, but I am not
obtaining sensible results. I'm thinking about trying the latter approach,
but I'd appreciate some confirmation and explanation on the differences
between the two approaches.

A follow up question: if I'm trying to perform DICS beamforming on my TFR
results, would I have to obtain a different spatial filter for each
frequency (assuming I use the second approach of analyzing multiple
individual frequencies)? I ask this because computing the spatial filter
requires the CSD matrix, which is a function of frequency.

Please let me know if there are any comments to be made. I appreciate any
help in advance!

Best,
Philip
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220317/2a8d9c9d/attachment.htm>