[Eeglablist] Fwd: Study Epoch consistency with single epoch

[Tarik S Bel-Bahar]

Hi Dave, a few thoughts below. If you come up with a good solution, thanks
for sharing it with the list so that other users can learn from your
experiences.

***********Notes for Dave
*My understanding is that eeglab developers setup study to expect at least
2 or more trials per condition, but you may hear back from them with
specifics.

*consider eliminating from your analyses the conditions were there are only
a few or one epochs

*note that comparing conditions that each have a different number of epochs
is not considered a good thing in eeg research, wherein one needs to at
least (statistically) control for the variable number of epochs one has
across the different conditions. It's defendable if you have, say
129,135,152,142 epochs across 4 conditions; but if you're comparing, say
1,5,8,20 epochs across 4condition, that's problematic, at least due to
noise/variability with low epoch counts and single epochs.

*in eeg research world, it's usual to have 40+ epochs per condition, though
for some well-established ERPs 15 epochs may be enough. This relates to the
general idea that one needs to average across a number of epochs (aka
trials) to get a clear/good event-related signal, and thus reduce the noise
across the epochs. This depends to some degree on the phenomena under
study, and the specific data collection protocol being used.

*in your case, you may need to go outside of study and get EEG metrics (and
average them up and run stats on them) from the individual single-subject
files. Though I don't recommend it, doing so would at least allow you to
get an estimate of your chosen EEG metrics for your one-epoch conditions
(regardless of the validity issue with one or few epochs).

*if you haven't had a chance to, consider reviewing methods and analyses in
previous published eeg articles that are similar to your eeg study, and
check out what their usual epoch counts are, and whether any
analyze/include conditions where there are just one epoch.

*there is a burgeoning literature one can check out on single-trial metrics
where one may find some useful techniques

*re wrangling with STUDY: One could try testing something out with the
eeglab tutorial data (the tutorial data that has multiple .set files with
epochs in each file, which can be loaded up into a study - and then delete
all the epochs except one from one single-subject file for one condition,
and then see if study works - this would be evidence that study can deal
with single-subject files that have one epoch)













On Tue, Jan 29, 2019 at 3:22 PM David Ryan <ryand1 at goldmail.etsu.edu> wrote:

> Hello,
>
> In my study I have several conditions (A, B,C…) and the epochs get
> organized into categories based on response (e.g., Condition A-Correct,
> Condition A-Incorrect, Condition B-Correct, Condition B-Incorrect, and so
> on…). This results with several single response epochs for certain
> categories. When I load this into STUDY it says “Epoch Consistency = no”
> and leads to matrix dimension errors for the ERSP.
>
>
>
> I have read through the Eeglablist and have found were several people want
> to analyze a single epoch as an epoch, not as continuous EEG - the EEGLAB
> default. The consistent response I have read is “ If there is one epoch, it
> cannot be recognized as epoched”. My question is why?
>
>
>
> How does a single epoch from a single subject in a condition with other
> multiple epochs from other subjects, all with the same epoch length not
> meet the criteria of epoch consistency? I know that EEGLAB reads a single
> epoch as continuous, but is there no way to change/override this assumption?
>
>
>
> In my own naïve attempt to address this issue, I tried to make single
> epochs “seem” the same in the STUDY structure by copying and formatting the
> ALLEEG.event information over to the ALLEEG.epoch. This still leaves the
> issue of the ALLEEG.trials =1.
>
>
>
> Any more ideas to address this issue?
>
>
>
> Thanks,
>
> Dave
>
>
>
>
>
>
>
> David Ryan, Ph.D.
>
> Research Health Science Specialist
>
> Hearing & Balance Research Program
>
> James H. Quillen VAMC (621)
>
> Research Service (151)
>
> Mountain Home, TN 37684
>
> (423)-926-1171 Ext. 7575
>
> [image: HBRH]
>
>
> --
>
> David Ryan, Ph.D.
>
> Research Health Specialist
>
> Mountain Home VA Health Care System (621)
>
> Auditory Vestibular Research Enhancement Award Program (REAP)
>
> Research Service
>
> (423)-926-1171 Ext. 7575
>
> www.avreap.research.va.gov
>
> _______________________________________________
> 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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20190130/624aaccd/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 16617 bytes
Desc: not available
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20190130/624aaccd/attachment.png>