[Eeglablist] LIMO First-level analysis

Fri May 13 00:19:39 PDT 2022

Dear Dr. Pernet,

Thank you for your prompt reply. I executed the methods you suggested and I
am still yielding the same problem. However, I was able to debug the
problem (although I am still not sure how to fix it)~

So the LIMO structure created for each subject has another structure within
it called LIMO.data. In this LIMO.data, I noticed that subjects for whom
all 3 jobs are successfully run have the same number of trials in the Cat
(Categorical variable) and Cont (Continuous variable~ my covariates) as
they were in the original EEG.data structure. However, for those
unsuccessful subjects, in the LIMO.data both Cat and Cont variables have
equal number of trials (100) but this number doesn't correspond to the
subject-specific total trials in their EEG.data structure.

For instance, sub-2 may have 45 trials in the EEG.data structure but in the
LIMO.data it will have 100x1 Cat and 100x6 Cont matrices.
This 100 digit is actually the total number of trials in my task. However,
after artifact rejection many subjects don't retain all the trials (though
there are some subjects with all 100 trials intact). Importantly, all
subjects for which LIMO can successfully complete all 3 jobs don't
necessarily have all 100 trials intact (e.g. sub-1 has 75 trials, sub-3 has
100 trials, sub-9 has 94 trials but they all have all 3 jobs successfully
completed).

I wonder why LIMO creates additional trials for some of these subjects
while for the others it accepts the original total number of trials/epochs.
Because as I mentioned in my previous email, all individual dataset were
batch preprocessed where within each individual dataset the total trial
numbers correspond to the total number of epochs. Your help is greatly
appreciated!

Regards,

Mo

On Fri, May 13, 2022 at 12:19 PM Dr Cyril, Pernet <wamcyril at gmail.com>
wrote:

> Hi Varma,
>
> mmh ok this might be an 'OLS' vs 'WLS' issue we solved recently(ish) but
> the LIMO plug-in hasn't been updated yet - so can you try rerunning the
> pipeline after upgrading
>
> -- download the code from master
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Ftools&d=DwIFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=90qyoC0YWWGweSn44AN9-qntnvrGSdTvRPG1f-mNA4mMK7WbJiG2bjzT4AVUIbGR&s=ApDCdvlHCvrIcBTNfPjEtHvzHx3jduPaJtADsdsuZ2s&e= 
>
> -- you can re-run the pipeline instead of using the STUDY and
> recomputing it all (ie it will automatically redo the failing subjects
> only --
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Ftools_wiki_debugging-2D1st-2Dlevel-2Dstats&d=DwIFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=90qyoC0YWWGweSn44AN9-qntnvrGSdTvRPG1f-mNA4mMK7WbJiG2bjzT4AVUIbGR&s=xglhALfPd92MSxdmJnh6XFuGmrxJuSJJP21vHMJm55Q&e= 
> )
>
> let me know how this goes
>
> thx
>
> cyril
>
>
>
> On 12/05/2022 18.29, VARMA, MOHITH MUKUND wrote:
> > Dear all,
> >
> > I am using the LIMO toolbox on EEGLAB v2021.1 to run regression at the
> > trial level ERP with one categorical variable and three continuous
> > variables. I followed the tutorial found here:
> >
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Fmeeg_wiki_12.-2DRegression-2Dat-2Dthe-2Dtrial-2Dlevel&d=DwIFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=Ra23jE11EDE2IH-KSLqT_vNQ2Ek5tsdA7Rz_ujHLOmYbVSf5oDOHV9wJ5S4Elexd&s=Zsnfs5u8zz8X-cCKt4iCKWFodPFPcYioHQg40bGmcSw&e=
> > <
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Fmeeg_wiki_12.-2DRegression-2Dat-2Dthe-2Dtrial-2Dlevel&d=DwMFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=JZ0vv-p51yYPbd2RRXVtX4epJP1LstD3tVtfN3EwI4ncPcc7rb-WZCifAtCyK6iO&s=wn7rpEpat1l3n266QMd3Tb32R3e0GEXtEIPv1krTci8&e=
> >
> >
> > The problem I am encountering is that LIMO is able to successfully run
> all
> > 3 jobs (import, design and glm) only for some subjects' data as shown in
> > the report. For the remaining subjects, LIMO is able to successfully
> import
> > the data but fails at the design and glm step.
> > I am not sure why only for some subjects LIMO cannot successfully run the
> > first-level analysis. All the dataset were batch preprocessed (ICA
> > correction, artifact rejection, epochs extracted without baseline
> > correction: as shown here:
> >
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Fmeeg_wiki_1.-2DPreprocessing&d=DwIFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=Ra23jE11EDE2IH-KSLqT_vNQ2Ek5tsdA7Rz_ujHLOmYbVSf5oDOHV9wJ5S4Elexd&s=xhw3_kkYlJtSe-vsbehW6gTVs8L44pJ97RaaPweYD4Q&e=
> > <
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_LIMO-2DEEG-2DToolbox_limo-5Fmeeg_wiki_1.-2DPreprocessing&d=DwMFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=JZ0vv-p51yYPbd2RRXVtX4epJP1LstD3tVtfN3EwI4ncPcc7rb-WZCifAtCyK6iO&s=AFjAvILBuHa-Z_-POSkrCnzKg4AFK3eiCVaWLludezw&e=
> >
> > ).
> > In each individual dataset trial numbers (EEG.trial) are equal to the
> total
> > number of epochs and events (EEG.epoch and EEG.event), so it is quite
> > strange that I am encountering this problem.
> >
> > Help from the experts will be greatly appreciated! Please feel free to
> let
> > me know if I need to share my batch script for your reference in the
> > follow-up.
> >
> > Regards,
> >
> > Mo
> > --
> > Mohith M. Varma, AFHEA
> > Research Assistant
> >
> > Department of Management
> > School of Business
> > Hong Kong Baptist University
> > Email: mohith96 at hkbu.edu.hk <mohith96 at connect.hku.hk>
> > _______________________________________________
> > 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
>
> --
> Dr Cyril Pernet, PhD, OHBM fellow, SSI fellow
> Neurobiology Research Unit,
> Building 8057, Blegdamsvej 9
> Copenhagen University Hospital, Rigshospitalet
> DK-2100 Copenhagen, Denmark
>
> wamcyril at gmail.com
> https://urldefense.proofpoint.com/v2/url?u=https-3A__cpernet.github.io_&d=DwIFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=90qyoC0YWWGweSn44AN9-qntnvrGSdTvRPG1f-mNA4mMK7WbJiG2bjzT4AVUIbGR&s=sLLDmVnIpouN2dCzAlejibUMWdkT19-bL9Lbb7vsuZ0&e= 
> https://urldefense.proofpoint.com/v2/url?u=https-3A__orcid.org_0000-2D0003-2D4010-2D4632&d=DwIFaQ&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=90qyoC0YWWGweSn44AN9-qntnvrGSdTvRPG1f-mNA4mMK7WbJiG2bjzT4AVUIbGR&s=JJnff2z30kcZIfMBEPwKhDRDFIXjYu6Y-0E7uIx0Q-M&e= 
>
>

-- 
Mohith M. Varma, AFHEA
Research Assistant

Department of Management
School of Business
Hong Kong Baptist University
Email: mohith96@hkbu.edu.hk <mohith96 at connect.hku.hk>