[Eeglablist] Study Design questions

Fiona Mitchell Baumer fbaumer at stanford.edu
Tue Jul 27 13:47:25 PDT 2021


I am trying to analyze some clinically gathered EEG data, focusing on epileptiform abnormalities. I have questions about how to use the STUDY function and cannot find information in the documentation. Would greatly appreciate your insight or if you could point me to more detailed documentation for these scenarios.

My main questions are:

  1.  The study function does not automatically populate data from the same subject as “paired” (i.e. says pairing “off” under the design function). Should I expect EEGlab to recognize paired data automatically if there is a subject with various conditions loaded? Should I set this within the Matlab editor or should I use something else in the GUI to do this? (only thing of note is that my subject names are a bit unusual as I retained data keys, i.e. AA12345A_1-1+; so I’m not sure if these characters interefere?)

  1.  I have 2 groups (~57 subjects with epilepsy and 57 controls). Subjects with epilepsy have up to 6 conditions and controls have up to 2 conditions. For some analyses, I want to compare conditions within a group. For other analyses, I want to compare the same condition between the 2 groups. Each condition for the 2 groups is loaded as its own set file. I set up 12 study designs, each with 2 categorical variables (condition and group).
     *   For the designs in which I wanted to compare a specific condition within the epilepsy group, I chose the 2 conditions of interest for the “condition” variable and picked only the epilepsy group for the group variable.
     *   For the designs in which I wanted to compare one condition between 2 groups, I did the opposite (picked 1 condition for the condition variable; picked both epilepsy and controls for the group variable).
     *   I don’t believe that this is excluding people appropriately however as when I look at STUDY.design.cases, for each of the 12 designs, it lists all 114 of my subjects for every design. The data also look odd when plotted. Is there some other way I should be doing this so that it only includes the subjects that have the conditions of interest/are from the group of interest?

Thank you all very much!

Fiona Baumer, MD
Assistant Professor in Child Neurology
Lucile Packard Children’s Hospital
750 Welch Rd, Suite 317
Palo Alto, CA 94304

More information about the eeglablist mailing list