[FieldTrip] timelock analysis for source analysis time axis

[Eelke Spaak]

Hi Xavier,

I would change the time axes of the chopped up segments to be
identical *before* the call to ft_timelockanalsis, e.g. with
data.time(:) = {0:1/fsample:2}; .

Cheers,
Eelke

On Mon, 29 Jul 2019 at 01:26, Xavier Vrijdag <x.vrijdag at auckland.ac.nz> wrote:
>
> Hello Fieldtrip experts,
>
>
>
> I am using fieldtrip to do LCMV source analysis. I have a dataset of baseline and 3 exposures to nitrous oxide of which I have 1 minute of EEG recording each. For the analysis I split the 60 second recording into 2 second trials using redefine trial. This creates a struct with the 30 trials and 30 time axis each trial having its own time axis 0:2 2:4 etc.  When calling the timelock analysis for the covariance matrix (with keep trials on) on this dataset I get a struct:
>
>       time: [1×15360 double]
>
>      label: {32×1 cell}
>
>      trial: [55×32×15360 double]
>
>     dimord: 'rpt_chan_time'
>
>        cov: [55×32×32 double]
>
>        cfg: [1×1 struct]
>
>
>
> This struct has 1 time axis from 0:60 seconds (fsample=256), with all the trials having their data on the correct time padded with NaNs before and after.
>
>
>
> Is this structure useful (keeping the original time axis 0:60) for the source analysis, or is it better to redefine the time axis for all trials to be 0:2? Is there a Fieldtrip function/setting to achieve this? The downside I currently experience is that the data struct is getting very big, causing an out of memory error.
>
>
>
> Thank you so much for helping me.
>
>
>
> Xavier Vrijdag
>
> PhD student
>
> University of Auckland
>
> _______________________________________________
> fieldtrip mailing list
> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> https://doi.org/10.1371/journal.pcbi.1002202