<div dir="ltr"><div class="gmail_default" style="color:#333399">Okay thanks Mahsa! remember you can run an ICA with 1 Hz highpass and then apply the ICA weights to continuous pre-ICA data that are filtered differently.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">Also, make sure you are appropriately correcting for rank.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">It's probably best to stay overall with the components that are truly neural. regarding whether to keep the mixed ones or not, you may review existing literature that works with similar data and protocols as yours. </div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">yes, there is a tradeoff regarding keeping or dropping mixed components that one has to deal with and understand the pros/cons.. Your "other" components if they don't have clear neural information and/or spectral peak, are probably candidates for removal.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Nov 13, 2017 at 9:42 PM, mahsa shalchy <span dir="ltr"><<a href="mailto:mahsa.shalchy@gmail.com" target="_blank">mahsa.shalchy@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Dear Tarik,<div><br></div><div>Thank you so much for your explanation. I had tried 1Hz HP cut-off and it did not improve my ICA results and since I am also interested in early components such as P1, I decided to go with 0.1 Hz. Also, I am visually rejecting my epochs before feeding them to ICA but unfortunately I forgot to mention. Using Luca's IC labeling and SASICA I'm able to keep only neural components, however, mixed or "other" components are usually the highest ICs ( IC3 to IC9) and by discarding them I am not only losing great amount of information but also ending up with low SNR.</div><div><br></div><div>Many thanks,</div><div>Mahsa</div></div><div class="HOEnZb"><div class="h5"><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Nov 13, 2017 at 11:44 AM, Tarik S Bel-Bahar <span dir="ltr"><<a href="mailto:tarikbelbahar@gmail.com" target="_blank">tarikbelbahar@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div class="gmail_default" style="color:#333399">Hello Mahsa,</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">You may want to use a 1 Hz bottom for your filter to get better ICA.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">It's probably best to remove any components that are dominated by eye blinks. </div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">I would recommend just sticking with analysis of just the IC components that are valid/neural, and ignoring/discounting the rest.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">If you are rejecting IC components and rebuilding the eeg data: Some researchers try to be extra cautious and they try not to remove ICs that are "mixed", so as not to inadvertantly remove neural signal.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">Note that ICA is not a perfect method, and that there are often ICs that have both artifact and neural signal.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">See Luca's reaching site to get info and training on selecting/reviewing ICs, including a "practice" option that provides feedback about your classification. At minimum, I recommend that students do about 200 classifications, review with an ICA-EEG expert, and then do another 200-300 classifications.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">Remember that you also have to do cleaning before ICA, which does not seem to be one of the steps you listed.</div><div class="gmail_default" style="color:#333399">If you haven't had a chance to yet, please be sure to review the eeglab tutorial steps, practice with with eeglab tutorial data, and check out Makoto's pipeline suggestions.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399">If you haven't yet, you should check out the various ICA classification plugins, such as MARA and SASICA, which should help you make a decision on your ICs.</div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_default" style="color:#333399"><br></div><div class="gmail_extra"><br><div class="gmail_quote"><div><div class="m_-3056107865171029360h5">On Thu, Nov 9, 2017 at 3:54 PM, mahsa shalchy <span dir="ltr"><<a href="mailto:mahsa.shalchy@gmail.com" target="_blank">mahsa.shalchy@gmail.com</a>></span> wrote:<br></div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="m_-3056107865171029360h5"><div dir="ltr">Dear Experts,<div><br></div><div>I am trying to correct EEG data blinks and saccades using ICA, however, in some cases (i.e. when the subject is double blinking), ICA finds multiple eye related components. </div><div><br></div><div>The problem is that except for the main blink component, other eye contaminated ICs, have neural information and I can not easily discard them ( please find an image of ICs here:</div><div><a href="https://www.dropbox.com/sh/qjz2v3ke6zdr1j1/AABoSzG4s6ZyAlUz_vYC8slsa?dl=0" target="_blank">https://www.dropbox.com/sh/qjz<wbr>2v3ke6zdr1j1/AABoSzG4s6ZyAlUz_<wbr>vYC8slsa?dl=0</a>)</div><div><br></div><div>I tried using PCA before ICA but it didn't seem to work. </div><div><br></div><div>For the preprocessing I did the following steps before applying ICA:</div><div>Resampling, Filtering (0.1-40), re-referencing to average mastoids,epoching.</div><div><br></div><div>I highly appreciate your help on this problem.</div><div><br></div><div>Many thanks,</div><div>Mahsa</div></div>
<br></div></div>______________________________<wbr>_________________<br>
Eeglablist page: <a href="http://sccn.ucsd.edu/eeglab/eeglabmail.html" rel="noreferrer" target="_blank">http://sccn.ucsd.edu/eeglab/ee<wbr>glabmail.html</a><br>
To unsubscribe, send an empty email to <a href="mailto:eeglablist-unsubscribe@sccn.ucsd.edu" target="_blank">eeglablist-unsubscribe@sccn.uc<wbr>sd.edu</a><br>
For digest mode, send an email with the subject "set digest mime" to <a href="mailto:eeglablist-request@sccn.ucsd.edu" target="_blank">eeglablist-request@sccn.ucsd.e<wbr>du</a><br></blockquote></div><br></div></div>
</blockquote></div><br></div>
</div></div></blockquote></div><br></div>