[Eeglablist] epoch length inquiry for short trials

[Makoto Miyakoshi]

Dear Brian,

I agree with Scott.
Wavelet transform/STFT uses a sliding window. If your default 3-cycle at
3-Hz (1116 ms long) window is applied to produce ERSP/ITC data for every 4
ms (assuming 250 Hz sampling), each sliding window is 1112 ms overlapped
with the adjacent one! So the outcome is already very redundant in that way
so no worry about another redundancy. You don't lose anything except for
the computational time, and you are not biasing the data in any
particular direction. There is no problem if redundancy is not a sin in the
given situation!

Makoto

On Mon, Jan 3, 2022 at 7:15 AM Kavanaugh, Brian via eeglablist <
eeglablist at sccn.ucsd.edu> wrote:

> Hi Scott,
>
> Wonderful, thank you for your guidance!
>
> Brian
>
> On Mon, Jan 3, 2022 at 12:06 AM Scott Makeig <smakeig at gmail.com> wrote:
>
> > Brian -
> >
> > There is no problem with extracting overlapping epochs if you keep in
> mind
> > which parts overlap.  If you are interested in measures in the
> > time/frequency domain (ERSP, ITC, coherence and cross-coherence, etc.),
> > then it is necessary to use a long enough trial epoch to include, e.g.,
> at
> > least one time window wholly within the trial 'baseline' period before
> the
> > first latency of analysis interest (to enable within-trial
> > before/during/after comparisons), and at least one window centered at the
> > last latency of analysis interest. To measure down to 3 Hz - using a
> > 3-cycle window to improve spectral resolution - a -1-s to 2-s window
> length
> > is sufficient.
> >
> > Scott Makeig
> >
> > On Sun, Jan 2, 2022 at 11:55 PM Kavanaugh, Brian via eeglablist <
> > eeglablist at sccn.ucsd.edu> wrote:
> >
> >> Hi everyone,
> >>
> >> Our kids go/nogo task has shorter trials (500 ms stimuli, 900 ms
> response
> >> window [or until response], 200 ms inter-trial interval) than the -1 to
> 2
> >> second epoch time length. Particularly as we are interested in midline
> >> theta activity.
> >>
> >> I'm sure this has been explained before, but I struggled to find it in
> the
> >> archives. I apologize in advance.
> >>
> >> If I use the -1 to 2 sec epoch in eeglab, won't these epochs overlap
> >> trials? Making inter-trial intervals longer would seem to solve this
> >> issue,
> >> but now that data has been collected... Any advice on optimal epoching?
> >>
> >> Thank you!!!
> >>
> >> Brian
> >>
> >>
> >> Brian Kavanaugh, Psy.D., ABPP
> >>
> >> Board Certified in Clinical Neuropsychology, Pediatric Subspecialist
> >>
> >> Pediatric Neuropsychologist, E. P. Bradley Hospital
> >>
> >> Assistant Professor (Research), Warren Alpert Medical School of Brown
> >> University
> >> _______________________________________________
> >> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> >> To unsubscribe, send an empty email to
> >> eeglablist-unsubscribe at sccn.ucsd.edu
> >> For digest mode, send an email with the subject "set digest mime" to
> >> eeglablist-request at sccn.ucsd.edu
> >>
> >
> >
> > --
> > Scott Makeig, Research Scientist and Director, Swartz Center for
> > Computational Neuroscience, Institute for Neural Computation, University
> of
> > California San Diego, La Jolla CA 92093-0559,
> http://sccn.ucsd.edu/~scott
> >
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>