Skip to content

Unfold 1.1
Compare
Choose a tag to compare
@behinger behinger released this 27 Mar 08:58
· 70 commits to master since this release
1.1
590ec76
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

Unfold 1.1

Features

single trial erpimages

grafik

Calculate ERPimages as seen in our publications. These erpimages can be of raw data, modelled data or of the residuals. Options to additionally include residual activity into the modelled is available. In addition one can remove/keep predictors from the modelled/residual output. This is helpful to e.g. simulate what it would look like to keep eventA, but remove eventB. Or to remove e.g. the effect of saccadeAmplitude but keep the effect of FixationPosition etc. See the new tutorial for more usage cases

Temporal response functions (TRFs)

grafik

TRFs allow to estimate brain responses from time-continuous predictors. E.g. if you record an audiostream concurrently with the EEG signal, you could include it as a timeexpanded predictorset. You could also use it for features of audio, pupil, continuous stimulation streams etc. See the tutorial for a usage example.

R² and partial R² (crossvalidated)

calculate explained variability of the model. Total R² over all modeled data and partial R² (totalR² - R²withoutPredictor), both with or without crossvalidation. Currently it is not possible to calculate R² for single time-points, but only for the complete modelled data.

uf_addmarginal

Addmarginal now supports marginal effect at the mean (MEM) and average marginal effect (AME). In short it is the difference of E[f(x)] and f(E[x]). This is the same for continuous variables but not for splines. Starting from Unfold 2.0 we will set the default to AME. For backward compatibility we will keep it by default at MEM

3 new tutorials

Tutorials on the peerJ paper, on erpimages, on temporal response functions

Bugfixes

  • Major: uf_continuousArtifactDetect.m had a bug inherited from ERPlab. For artefact detection only the last channel was used!! This is now fixed.
  • init_unfold now uses eeglab redraw instead of eeglab, not overwriting an already loaded ÈEG structure.
  • interactions with more than two levels had a bug while renaming, only if they occured in multiple events.
  • several smaller bugfixes
Assets 3
Loading