Unfold Documentation
Documentation is currently written.
Best to start with installation, then go for the mass-univariate approach lm_mu.md, which should be familiar to you if you did ERPs before. Then the overlap-correction lm_overlap.md, LMM (to-be-done) and non-linear (to-be-done) would be good stations.
In case you want to understand the tools better, check out our explanations.
Once you are familiar with the tools, check out further how-to guides for specific applications.
In case you want to understand the toolbox better, we plan to offer technical references. This includes Benchmarks & Explorations.
Summary
There are four different model types currently "fitable"
- Timeexpansion No, Mixed No :
fit(UnfoldModel,Dict(Any=>(f,-0.1:0.01:0.5)),evts,data_epoch)
- Timeexpansion Yes, Mixed No :
fit(UnfoldModel,Dict(Any=>(f,basisfunction)),evts,data)
- Timeexpansion No, Mixed Yes :
fit(UnfoldModel,Dict(Any=>(fLMM,-0.1:0.01:0.5)),evts,data_epoch)
- Timeexpansion Yes, Mixed Yes:
fit(UnfoldModel,Dict(Any=>(fLMM,basisfunction)),evts,data)
With f = @formula 0~1+condition
fLMM = @formula 0~1+condition+(1|subject) + (1|item)
basisfunction = firbasis(τ=(-0.1,0.5),sfreq=100"))