[Eeglablist] Single trial connectivity analysis using SIFT

[Makoto Miyakoshi]

Dear Sampath,

What model order did you use?
If you have no idea what it is, see my example (taken from Tim Mullen's)
that automatically optimizes the model order.
https://urldefense.com/v3/__https://github.com/sccn/groupSIFT/blob/cd56e3166cfccd898420532e4ef6a981550205f9/pop_groupSIFT_runSiftBatch.m*L384__;Iw!!Mih3wA!ALuRYGfzunk3OAEmHVIfzr-U0s4zEPesMkrnz3JoOh7C3qySpWRJF13a-fghwTPEN6Vbm1wxirwjOG-b5recqNTtDgE$ 
For continuous data processing, your sliding window length must be MUCH
longer. Consider > 10-20 s.

Makoto

On Wed, Feb 8, 2023 at 5:26 PM Sampath Thoutanahalli Kapanaiah <
sampath.kapanaiah at uni-ulm.de> wrote:

> Dear Makoto,
>
> Thank you very much for the reply. I use a modified script from your
> example continuous script,  although I still get below error more often.
>
> Warning: The specified AR model is unstable at timeindex 1.
> > In tvarsim (line 112)
>   In est_consistency (line 78)
>   In est_checkMVARConsistency (line 135)
>   In est_validateMVAR (line 210)
>   In pop_est_validateMVAR (line 147)
>
> Do you recommend any particular parameter? I am currently using the below
> parameters.
>
> WindowLengthSec     = 3;
> WindowStepSizeSec  = 3;
> algorithm                   = 'ARfit';   %%%Vieira-Morf, ARfit
> windowingType          = 'rectwin'; %%%%'hamming' or 'blackmanharris'
> detrend_method       = 'linear';
> normalise_method    = {'time'};
> connmethods             = {'iCoh' 'ffDTF' 'GPDC' 'RPDC' 'GGC' 'dDTF08'};
> GUI_MODE                   = 'nogui';
> VERBOSITY_LEVEL       = 0;
>
>
> Thank you very much for your time
>
> Best Wishes,
> Sampath
>
>