[Eeglablist] Question about the statistical result of the main effects on ERSP data using EEGLAB GUI

[Park, Jinhan]

Dear Dr. Delorme, 

Thank you for letting me know I was right. 
Plus, I tried doing your recommendation that I use the version of codes you let me know. However, the condition main effect still looked like the group main effect with GUI. 

Also, I have one more question about the "std_stat.m" function. I have been developing codes using the EEGLAB functions for a group analysis. When I did do statistical analyses for ERSP, I have used the function. Then, once I got the outcomes from the functions, the function provided me with the following outcomes:
1. pcond: logical value for conditions (if a p-value is over .05, the value should be 1)
2. pgroup: logical value for groups (if a p-value is over .05, the value should be 1)
3. pinter: logical value for 2-way ANOVA (if a p-value is over .05, the value should be 1). This outcome has three cells which represent results from each f-test (A, B, AxB).
4. statscond
5. statsgroup
6. statsinter

According to the explanation of the "std_stat.m", the order of the "pinter" outcome should be as follows: a main effect of conditions -- a main effect of groups -- an interaction effect. 
My question is that my understanding is correct and that if I am right, can I switch the main effects of conditions and groups each other (because I think that main effects from GUI looked opposite; also, when I looked at the outcomes of the "pinter", it has the same problem like the GUI)? 

I am looking forward to your answer!
Thank you so much!

Best regards, 
Jinhan

-----Original Message-----
From: Arnaud Delorme <adelorme at ucsd.edu> 
Sent: Tuesday, November 8, 2022 7:24 PM
To: Park, Jinhan <jinhan.park at ufl.edu>
Cc: eeglablist at sccn.ucsd.edu
Subject: Re: [Eeglablist] Question about the statistical result of the main effects on ERSP data using EEGLAB GUI

[External Email]

Dear Jinhan,

Yes, you are right. This has been fixed last September.
The code on GitHub is up to day but it will not be released until Jan 2023. Can you check that the GitHub Version works for you.

https://urldefense.com/v3/__https://nam10.safelinks.protection.outlook.com/?url=https*3A*2F*2Fgithub.com*2Fsccn*2Feeglab*2Fcommit*2F345247b6023462f54baaf4075579da91d36b40d6&data=05*7C01*7Cjinhan.park*40ufl.edu*7C81cae32cd8f045d1911b08dac1e8a747*7C0d4da0f84a314d76ace60a62331e1b84*7C0*7C0*7C638035502249733602*7CUnknown*7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0*3D*7C3000*7C*7C*7C&sdata=fNK2iiS4NgqJH5F1rmiIQvoqwRmWN*2FQzVILhgd3Qht4*3D&reserved=0__;JSUlJSUlJSUlJSUlJSUlJSUlJSUlJSUl!!Mih3wA!D0F_1__caxvx-NVx6s14Pgtp8MWO5rJ5OxJffHv4TUQRLFOWaOl0jmllnyJtpPP7jmkx3CwNqWSfFBaig2-osyNcT5s$  

Arno

> On Nov 4, 2022, at 4:05 AM, Park, Jinhan <jinhan.park at ufl.edu> wrote:
>
> Dear all,
>
> At a cluster level, I am analyzing ERSP data of a two-way (2 conditions x 2 groups) ANOVA design using EEGLAB GUI.
> When I plotted the figures of the two-way ANOVA results, however, the figure that represented the main effect for the conditions seemed to be more for the main effect for groups, not the conditions.
>
> The procedures to get the ERSP figures using GUI:
>
>  1.  Create a STUDY.
>  2.  Load each subject and subject's condition datasets individually preprocessed.
>  3.  Go to "Study" and then Click "Select/Edit study design(s)".
>  4.  Set up STUDY design --> conditions (High-Gain vs. Low-Gain) X groups (Controls, Patients).
>  5.  Go to the "precompute component measures" --> Check the "ERSP" box.
>  6.  "PCA clustering (optional)" --> "Build preclustering array" --> Check the "ERSPs", "dipole locations" and "dipole orient" --> Hit the "OK"
>  7.  "PCA clustering (optional)" --> "cluster components" --> Use "Kmeans" and check the "Separate outliers (enter std.)" box --> Hit the "OK"
>  8.  Set up statistical methods --> "General statistical parameters" 
> --> Check the "Compute 1st independent variable statistics if any" and "Compute 2nd independent variable statistics if any" --> "Use EEGLAB statistics" --> "Permutation statistics" (n = 2000) and alpha = 0.05 (no correction for multiple comparison)  9.  Conduct "ERSP" for a cluster.
>
> As with the above procedures, I have tried doing it with other datasets (e.g., 2x3 two-way ANOVA design), but I am strongly sure the figure for the main effect of conditions looks like groups' main effect, not the conditions.
>
> I would sincerely appreciate it if you could let me know what is going on here.
>
> Best regards,
> Jinhan Park
>
>
> Jinhan Park | Ph.D. student
> College of Health and Human Performance University of Florida 
> Laboratory of Rehabilitation Neuroscience | lrnlab.org
> Phone: +1 (352) 328-6605
> Email: jinhan.park at ufl.edu<mailto:jinhan.park at ufl.edu>
>
> _______________________________________________
> Eeglablist page: 
> https://urldefense.com/v3/__https://nam10.safelinks.protection.outlook.com/?url=http*3A*2F*2Fsccn__;JSUl!!Mih3wA!D0F_1__caxvx-NVx6s14Pgtp8MWO5rJ5OxJffHv4TUQRLFOWaOl0jmllnyJtpPP7jmkx3CwNqWSfFBaig2-ovQNkK1g$  .
> ucsd.edu%2Feeglab%2Feeglabmail.html&data=05%7C01%7Cjinhan.park%40u
> fl.edu%7C81cae32cd8f045d1911b08dac1e8a747%7C0d4da0f84a314d76ace60a6233
> 1e1b84%7C0%7C0%7C638035502249733602%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC
> 4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%
> 7C&sdata=eeTJwTbpPW4gYZtdLW6IdeEmPl8CK5PeMXNpVu617%2Bs%3D&rese
> rved=0 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