Editing software issues

From Center for Cognitive Neuroscience
Revision as of 22:47, 28 January 2015 by Soconnell (talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Sanders, Andrew J. <AJSanders@mednet.ucla.edu> 11/24/14

Hi Samantha,

I am emailing because I remember you mentioning your lab is beginning to use Netstation 5.

Our collaborators in Boston have run into some issues with processing in Netstation 5 and are looking to see if others are experiencing similar issues.

Would you mind answering some of their questions, or putting them in contact with someone in your lab who may be more familiar with processing in NS5?

Thanks!!


On Nov 24, 2014, at 2:02 PM, Jeste, Shafali M.D. wrote:

NO and this is intentional :) However my collaborator Mark Cohen is, and I’m sure would be happy to talk with you. I’m cc’ing Andrew who can connect you to the right person in his lab.

SSJ

Shafali Spurling Jeste, MD Assistant Professor in Psychiatry and Biobehavioral Sciences UCLA David Geffen School of Medicine Director, Electrophysiology Core UCLA Center for Autism Research and Treatment 760 Westwood Plaza Semel Institute, Room 68-237B lab: 310-825-2761 clinic: 310-794-4008


On Nov 24, 2014, at 1:11 PM, Faja, Susan <Susan.Faja@childrens.harvard.edu> wrote:

Hi Shefali, Hope you are hanging in there with the U19 stuff. Thanks so much for all your help last week while you were out of the country!

Quick question: Has your lab been using Netstation 5 at all yet?

We're trying to set up a new system in my lab in Boston and have run into some issues with processing and just wanted to see if other users are experiencing similar things.

Thanks, Susan

-- Susan Faja, Ph.D. Labs of Cognitive Neuroscience | BCH3178 Boston Children’s Hospital | 1 Autumn Street | AU459 | Boston, MA 02215 617-919-4486 | susan.faja@childrens.harvard.edu



Samantha O'Connell <samroconnell@gmail.com> 11/25/14

to Andrew, Susan, Tessa, Shafali Hi Andrew,

We have not gotten to the stage of processing data as of yet - we've been having data collection timing issues with Psychtoolbox and NS5. Because we haven't collected data accurately yet, we haven't looked into processing.

If I can hear what kind of processing issues your collaborators have been having, I can address those to Mark Moran, our connect at EGI. It'll be good for us to know of these issues as well, for we're looking to do some processing using NetStation5 once we have it up and running.

Thanks, Sam


Jeste, Shafali M.D. <SJeste@mednet.ucla.edu> 11/26/14

to Susan, me, Andrew, Tessa UGH this sounds painful. However it is also a reality that we all will have to switch at some point, so working through the growing pains together will be helpful. I was just able to convince them to sell me another net amps 300, but they will only sell it with NS5 due to the computer/mac issues. However we are trying to scheme ways to work around that issue as I prefer to stick with 4.5 for now.

At some point a work group on this will probably be in order.

SSJ

Shafali Spurling Jeste, MD Assistant Professor in Psychiatry and Biobehavioral Sciences UCLA David Geffen School of Medicine Director, Electrophysiology Core UCLA Center for Autism Research and Treatment 760 Westwood Plaza Semel Institute, Room 68-237B lab: 310-825-2761 clinic: 310-794-4008


On Nov 25, 2014, at 11:18 PM, Faja, Susan <Susan.Faja@childrens.harvard.edu> wrote:

Hi Sam et al., We've been in touch with Mark Moran and EGI support a lot as we've had a whole slew of processing problems. I won't go into detail about our 'important but not critical list', but would be happy to send more if it's helpful. Some of the most critical issues are:

1) Temporal relations have been removed from the Segmentation Tool - this means that you can only segment to 1 thing and it just becomes 0 in your segment. In NS 4, we typically segmented to stim, but also pulled in criteria related to the accuracy and type of the previous trial. This is no longer a possibility. We also used a photocel during our acquisition to confirm timing and segmented to stim + DIN. Also no longer an option. Mark worked with us for about a week on a workaround. I ended up reprogramming. Mark tells us that EGI decided that NS users don't need temporal relations, but that he continues to lobby for it. If you would like to see this feature, please let him know we're not the only lab that used it.

2) The bad trials from artifact detection get reincorporated in later processing steps. (i.e. they don't stay bad). EGI is aware of this bug and telling us it'll be corrected in a later version.

3) NS averaging tool doesn't work - all the channels get marked as bad and zeroed out. Again, EGI is aware of this and telling us it'll be corrected in a later version.

4) NS 5 doesn't have stat extraction. EGI is aware of this and telling us it'll be corrected in a later version.

Re. 3 & 4, EGI recently told us that it's not been a problem for their other users because they are simply not processing in NS 5. This seemed bizarre to us. We are trying to unpack this by reaching out to labs that have EGI systems to see whether they are using 5 yet and what alternatives they've come up with. (We've also been on the EGI user group to learn about exporting to Matlab for ERPlab. The EEG signal exports, but it's unclear whether the NS tags can be exported--We're following up on some very cumbersome workarounds from John Richards who posted today).

Tessa, feel free to chime in if I have missed anything critical in this summary.

Thanks so much for letting us know where you are at with things and let us know if we can be helpful.

-Susan

p.s. Shefali, thanks! When I arrived in Boston, NA400 and NS5 were the only products shipping from Eugene. We get so sad when we see NS4 on Chuck's computers now. Oh the good old days...


Faja, Susan <Susan.Faja@childrens.harvard.edu> 11/26/14

to Shafali, me, Andrew, Tessa Thanks Shefali, I'm really happy to share my experiences if it can benefit others. (Sara W and Adam Naples have heard a fair amount and, of course, Chuck is in the loop). I think researchers using EGI will definitely need to stick together as we suspect they may be moving more toward clinical customers.

Have a wonderful Thanksgiving, Susan


Samantha O'Connell <samroconnell@gmail.com> 11/26/14

to Susan, Shafali, Andrew, Tessa Hi Susan,

Thanks for the input - I'll forward this over to my colleagues so they have a sense of what's going on, too. It's really too bad that there's been so many issues with NS5 on both the acquisition and processing ends, and I hope that they'll figure this all out soon. I really appreciate that you took the time to share. I'll be sure to let you know if anything comes up.

Thanks, Sam


Samantha O'Connell <samroconnell@gmail.com> 11/26/14

to Agatha, Andrew, Austin, Edward A colleague collaborating with Shafali in Boston has been working with NS5's processing software and it looks a little grim - see her outline below of the problems they've been encountering as well.

Thanks, Sam


Agatha Lenartowicz <alenarto@ucla.edu> 11/26/14

to me, Andrew, Austin, Edward Thanks very much Sam - this is all reason to not process in NS5 for us. Though of course any issues pertaining to collecting and extracting data are relevant.


Return to EGI main page