[FieldTrip] ft_freqanalysis using wavelets, problems with trialdef

Dominic Depke dominic.depke at rub.de
Wed Jun 18 15:58:25 CEST 2014


Dear all,

Actually I am analyzing EEG data containing epileptic seizure-like 
activity in Fieldtrip.
Data is read in continuous mode and segments of interest are visually 
identified using the databrowser and a trial-structure (cfg.trl) is made 
from the information in the 'cfg.artfctdef.visual.artifact' field.
After the trial-definition ft_preprocessing is called a second time with 
the option 'cfg.continuous' set to 'no'.
Now I tried to use ft_frequanalysis using wavelets to get information 
about power-spectral densities for every single trial and every single 
channel ( in terms of individual 'seizure' events) but the output 
structure is filled with NaNs.
I read the tutorial documentations and changed the values for cfg.toi 
and width but the result remained the same.

Further, my datastructure does not contain a .trialinfo field so how 
should the  trialdefinition look like (e.g. if you create cfg.trials)  
for use with ft_freqanalysis?

There is only a cfg.trl and a data.trial field in the datastructure. 
Moreover, my data does not contain any stimulus codes.

Best,
Dominic



More information about the fieldtrip mailing list