[FieldTrip] Diagonal elements in MVAR connectivity

[Lasha Ephremidze]

Hi,

Diagonal entries need not be zeros in MVAR, there might be nonzero
autoregressive coefficients.

Best regards,
Lasha Ephremidze

On Thu, Dec 8, 2022 at 7:28 PM Ioana Susnoschi-Luca (PGR) via fieldtrip <
fieldtrip at science.ru.nl> wrote:

> Hello!
>
> I am working on effective connectivity analysis with DTF and I am confused
> about the main diagonal elements - I expected my resulting DTF spectrum to
> have zero on the main diagonal, but this is not the case. Does it output
> the power spectrum of the fitted MVAR model instead?
>
> I would appreciate any insight into why this happens/if my thinking is
> correct. I reproduced my analysis steps (pretty much the same as FT
> connectivity tutorials) and rationale for diagonal values.
>
> Intuitively, DTF calculates the ratio between inflows from channel j -> i,
> and all inflows to i, hence I expected i->i to be zero. It uses the
> transfer matrix output from mvar frequency analysis, but the transf matrix
> i to i looks like the MVAR model's autospectrum for channel i, so I
> supposed the diagonal is a ratio between autospectra and sum of inflows to
> each channel.
>
> Thank you,
> Ina
>
> %% CODE
>
> % fit MVAR
> cfg          = [];
> cfg.order    = 19;
> cfg.method  = 'bsmart';
> cfg.demean  = 'yes';
> cfg.zscore    = 'yes';
> mvar_data   = ft_mvaranalysis(cfg, source_ts);
>
> % frequency analysis MVAR
> cfg         = [];
> cfg.method  = 'mvar';
> mfreq       = ft_freqanalysis(cfg, mvar_data);
>
> % DTF spectrum
> cfg         = [];
> cfg.method  = 'dtf';
> dtf         = ft_connectivityanalysis(cfg, mfreq);
>
> dtfPre_dir  = dtf.dtfspctrm;
>
> _______________________________________________
> fieldtrip mailing list
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__mailman.science.ru.nl_mailman_listinfo_fieldtrip&d=DwIGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=1ZP2y_-rhJRJWZpQyWZyCA&m=TBkuckNveyTw6Rp5Pmoyjz4D3sJkUG5ezPs_y3M7HK7M1hfMti5V42BlyQ8ues9r&s=xXCI2nY4eF9uCG2oxuO81oL6926kgkrQL2UXbaoJdZs&e=
>
> https://urldefense.proofpoint.com/v2/url?u=https-3A__doi.org_10.1371_journal.pcbi.1002202&d=DwIGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=1ZP2y_-rhJRJWZpQyWZyCA&m=TBkuckNveyTw6Rp5Pmoyjz4D3sJkUG5ezPs_y3M7HK7M1hfMti5V42BlyQ8ues9r&s=fWiisGuhK0uhIHBl1nFZLhpyqjT5ic4HUnUSwQVw_Jc&e=
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20221208/cf3c6455/attachment.htm>