[FieldTrip] Inferential statistics on connectivity data?

Katarina Slama slama at berkeley.edu
Mon Jan 4 02:22:43 CET 2016


Dear Fieldtrip community,

My name is Katarina Slama. I’m a graduate student in the Knight lab at UC
Berkeley, and I use ECoG to study visual attention.

I am currently examining connectivity in ECoG data using
ft_connectivityanalysis (cfg and data structures below). I have computed
partial coherence values for a network of electrodes. Now I would like to
assess the statistical significance of each of the edges by computing a
non-parametric equivalent of a one-sample t-test on the coherence values
over trials. I have two questions:

1. Is there a way to obtain a separate partial coherence value for each
trial? At present, the partial coherence structure that
ft_connectivityanalysis creates has the dimord “chan_chan_freq”.

2. Is there a fieldtrip function that is designed for computing statistics
on connectivity metrics? I have been advised to look into
ft_freqstatistics, but this function does not seem to be designed for the
output of ft_connectivityanalysis.

—
input structure (freq):
label: {'34'  '35'  '36'  '37'  '38'  '39'  '44'  '45'  '47'  '48'  '49'
 '50'  '51'  '52'  '53'  '54'  '55'  '56'  '57'  '62'  '63'  '64'  '76'
 '77'  '80'  '99'  '100'  '101'}
dimord: 'rpttap_chan_freq'
freq: [2.0000 3.0000 4.0000]
fourierspctrm: [220x28x3 double]
cumsumcnt: [220x1 double]
cumtapcnt: [220x1 double]
trialinfo: [220x1 double]
cfg: [1x1 struct]


configuration (cfg):
cfg = [];
cfg.method = 'coh';
cfg.partchannel = freq.label(1);
cfg.keeptrials = 'yes';

partcoh = ft_connectivityanalysis(cfg,freq);

output structure (partcoh)
label: {27x1 cell}
dimord: 'chan_chan_freq'
cohspctrm: [27x27x3 double]
freq: [2.0000 3.0000 4.0000]
cfg: [1x1 struct]
—

Thank you and best wishes,
Kata
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20160103/9b2608cc/attachment.html>


More information about the fieldtrip mailing list