LInear MOdeling
for Magneto-
and Electro-
Encephalography

Download

Download the zip file, clone or fork our master repository

Use with EEGLAB

LIMO tools were developped with EEGLAB functions for reading files and do some vizualization. Simply unpack the repository into the EEGLAB plugin folder. It can be called via STUDY using the EEGLAB developper version.

Stand Alone usage

All the statistics are performed internaly and therefore our tools can be used as stand alone. After setting the path to the toolbox, you can call statistical functions and vizualize results. Check the wiki on how to use each function for your analyses

FieldTrip wishes

The next frontiers ... there are no reasons why we could not integrate. We are actively looking for new contributors who want to make this happen.

Tools overview

Modern and Robust statistics for electrophysiology

Modelling

Flexible Models through a hiearchical approach

  • Categorical variables
  • Continuous variables
  • Between trial variance
  • Between subject variance

Estimating

Remove influence of outliers

  • OLS, WLS, IRLS
  • Robust estimators
  • Credible intervals

Testing

Use modern resampling techniques

  • Whole space analysis
  • FWER correction
  • Clustering
  • TFCE

How to contribute to the project

The prefered way to report bugs is to submit an issue on github, but emailing us is fine too. We'll try to adress your problem as soon as possible

Bug report

The fastest way to get acknowledged is to add your name to our contributor list

Get acknowledged

Found a bug and fixed it! submit the pull request to we can merge it - add your name on our contributor lists

Bug fixing

There are always new methods to use, maybe you designed one?. Get in touch to integrate it to the list of tools

Add your own algorithms

There many ways to contribute, including documentation and testing, check it out

Beyond bugs

Using LIMO tools - FAQ

one sample t-test, two-samples t-tests, paired t-tests, ANOVA, ANCOVA, repeated measures ANOVA, single and multiple regressions, and more to come.

Which tests are available?

For each subject, code each condition (e.g. 4 conditions in a 2x2 design). At the group level, compute a repeated measures ANOVA with N factors (eg. 2x2).

Analyzing factorial designs

These are statistical procedures that work for non-normal distributions and are this not affect by outliers.

What is robustness?

To compute clusters at the channel level, one must declare which channels are neighbours, i.e. which channels we want to see grouped together.

What are channel neighbours?

Results must be interpreted based on the statistical method used for inference, cell-wise or cluster-wise

Reporting results

Depending if you want to do a group analysis or single subject analyses, conditions can be declared as a single or multiple entries (factorial design per subject).

Define conditions