[FieldTrip] correlation | statfun

[Lars Meyer]

dear all,

on the new fieldtrip homepage, i find a short suggestion on how to try and run a correlation analysis on timelock data using an external regressor (e.g. an fmri-effect) and a single-rowed cfg.design containing this regressor. my questions now are: 

how does the data have to look? should i just produce an effect (= difference .avg containing the difference between two erps) for each participant, i.e. my input data is a 1 x participants struct?

is it correct that cfg.design is only the regressor (= one value per participant), and that cfg.uvar is empty?

if the above are correct: what is the proper statfun to run this type of statistic, i.e. a correlation between the individual effects and an external regressor, if i want to use the cluster statistics monte carlo approach?


thanks a lot,
l.


Lars Meyer
Max Planck Institute for Human Cognitive & Brain Sciences
Department of Neuropsychology
Stephanstrafle 1a
04103 Leipzig | Germany

Office | +49 341 99 40 22 66
Mail | lmeyer at cbs.mpg.de
Web | www.cbs.mpg.de/~lmeyer