[FieldTrip] Use of Granger Causality

[daniel.strahnen at uni-ulm.de]

Dear FieldTrip-Community,

 

I am seeking out for advice on the proper use of Granger Causality analysing
local field potentials from mice with different genotypes. 

The sampling rate is 10kHz, but I downsampled to 250Hz.

The general goal is to identify possible changes in directionality during a
10 min exploratory time and during specific task with a respective duration
of 2 seconds.

I have got the following questions:

1.	Should the parametric or the non-parametric approach be favoured? So
far I used the non-parametric approach by using FieldTrip and the parametric
approach by using the 

MVGC-toolbox by Barnett & Seth (2014). With the parametric approach the
optimal model order recommended by AIC or BIC is always very high so I am
really not sure if this

is the right way to go.

2.	After identifying possible peaks in Granger Causality (e.g. in the
Theta range), is it valid to compare them statistically using ANOVA between
different cohorts of mice? Or are

other statistical tests needed like a permutation procedure?

 

Thank you in advance and best regards

 

Daniel

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20190803/5bcfbdad/attachment-0002.html>