strange TFR effects

Bernhard Staresina bps231 at NYU.EDU
Sun Feb 7 19:35:30 CET 2010

Dear Fieldtrip experts,

I've come across a strange effect during the calculation of TFRs using
multitaper convolution. When I put in the raw data (no preprocessing at
all), I get TFRs that look very weird. However, if I use a bandpass filter
(preproc_bandpassfilter from 1 to 150 Hz with default settings) just prior
to calculating the TFRs, the results look perfectly normal. I'm attaching
figures of both versions so you can see the difference. The settings for the
TFR calculation are identical for both cases:

cfg              	= [];      	= 'all';
cfg.output       	= 'pow';
cfg.pad          	= 'maxperlen';
cfg.keeptrials   	= 'yes';
cfg.toi          	= -0.75:.01:2.0;
cfg.method       	= 'mtmconvol';
cfg.foi          	= 2:1:120;
cfg.t_ftimwin 	= ones(length(cfg.foi),1);
cfg.t_ftimwin(cfg.foi < 30)   = .500;
cfg.t_ftimwin(cfg.foi >= 30) = .200;
cfg.tapsmofrq    = linspace(2,20,length(cfg.foi)); 
cfg.keeptapers   = 'no';

The plots represent relative power changes, calculated by subtracting and
dividing through the power from -.450 to -.250 ms before stimulus onset.

Do you have any clue what's the reason for this difference?

Thanks for your help,

The aim of this list is to facilitate the discussion between users of the FieldTrip  toolbox, to share experiences and to discuss new ideas for MEG and EEG analysis. See also and
-------------- next part --------------
A non-text attachment was scrubbed...
Name: A_unfiltered.jpg
Type: image/jpeg
Size: 35829 bytes
Desc: not available
URL: <>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: B_bp_filtered.jpg
Type: image/jpeg
Size: 28425 bytes
Desc: not available
URL: <>

More information about the fieldtrip mailing list