Hackpads are smart collaborative documents. .

Feng Zhou

1029 days ago
Unfiled. Edited by Marianne Reddan , Leonie Koban , Feng Zhou 1029 days ago
Progress Updates:
Marianne R 7/31: Single trial models
 
  • preproc --> options to add: no slice timing correction, successive diff movies
DSGN files
  • add single trial model struc to the example DSGN file
 
  • confirm that mean imgs from each run look like brains
 
Variance Inflation Factor
Motivation: credit assignment issue (how do you distribute credit for success among the many decision that may have been involved in producing it?)
 
 
  • SINGLE TRIAL MODELS
SPECIFYING SINGLE TRIAL MODELS IN YOUR DSGN FILE
Motivation: useful for running multilevel mediation and looking at single trial effects of pain, etc
  • Scripts & Files You Will Need For This Section
  • DSGN file
Leonie K
  • canlab_glm_subject_levels
  1. Go into your first level model DSGN file and add this after your DSGN.conditions section. This references your conditions. You can indicate (with a 1) if you want that condition to be modeled as single trials or not. Easy!
Marianne R
  • %% SINGLE TRIAL
Leonie K
  • DSGN.singletrials{1} = {1 1 0 0};
Marianne R
  • %% this specifies which conditions will be modeled as single trials, in the order that you listed your conditions
Leonie K
  • %% e,g, in this case the first 2 conditions would be modeled as single trials, but not the third and fourth condition
  1. You do not need to specify contrasts in your script (usually not useful on the first level for single trial models)
Marianne R
  1. Now you are ready to run your models (see the First Level Models section of this document for more information).
 
Issues and Concerns
  • Watch Your VIFS
  • High VIFs (greater than 5 or 2.5, depending on whom you ask ;) ) can be harmful to your single trial model results and mediation results 
  • Rapid Event Related Designs
Use what works best for your design:
If your ITI is very short, you might have lots of overlap between trials, and high VIFs. One way to get over this would be to use a modeling scheme described by Mumford (2012). Figure is from there.
 
  • MEDIATION TOOLBOX
FIRST PHASE
Motivation
  • Scripts & Files You Will Need For This Section
  • mediation_brain_multilevel
  • every cell is one subject, and each have M (betas), X (e.g. -1/1 trial type), & Y (e.g. pain rating)
  • Steps
 
  • SVN: COMMITTING CHANGES TO THE REPOSITORY
UPDATING YOUR LOCAL COPY
Motivation:  You want to update your local copy of the repository
  • Steps
  1. open terminal
  1. cd to your repos folder:
Marianne R
  • cd ~/CANLabRepos
  1. update
  • svn update
 
LOOKING AT THE CHANGES YOU'VE MADE
Motivation:  You edited your local scripts, maybe you want to share these edits, maybe you don't want to yet. This lets you see what you've done.
  • Steps
  1. open terminal
  1. cd to your repos folder:
  • cd ~/CANLabRepos
  1. check what you've changed... this will print out all the files you've edited to
  • svn status
  1. You don't remember what you changed in one of the files listed? Let svn tell you, using 'diff' then the full path to the file
  • svn diff trunk/SCN_Core_Support/Visualization_functions/barplot_columns3.m
  1. You want to go back to the way things were? Use 'revert' and give the full path to the file
  • svn revert trunk/SCN_Core_Support/Visualization_functions/barplot_columns3.m
  1. You think your edits are golden and should be shared with everyone? Use 'commit' and give a little message about what you have changed. Important: make sure you check your status. You will commit everything you have changed, so revert things you do not want to commit.
  • svn commit -m "Added info on single trial models to the README -marianne"
 
 
 
------------SCRATCH PAD-----------
 
  • TOPIC
FIRST PHASE
Motivation
 

Contact Support



Please check out our How-to Guide and FAQ first to see if your question is already answered! :)

If you have a feature request, please add it to this pad. Thanks!


Log in