[FieldTrip] Cluster statistics on functional connectivity

[Schoffelen, J.M. (Jan Mathijs)]

Hi Julian,

Fieldtrip can only automatically define ’neighbours’ in node-space, not in edge-space. That is, providing a specification of cfg.neighbours (see ft_prepare_neighbours etc.), the lower-level function that deals with the spatial clustering can create a boolean adjacency matrix (#nodes x #nodes). This adjacency matrix is implicitly stored in cfg.connectivity. This works only if the input data is defined on the nodes.

In your case, where the input data is defined on edges (i.e. channel combinations), it should be theoretically possible to add a cfg.connectivity adjacency matrix, which is defined on the edges. Yet, you need to define this matrix yourself by hand (you can grep the code base for cfg.connectivity to get an idea of the relevant code). The crucial question there is that you need to define when two edges are defined as neighbours. Another thing to keep in mind is that the order of the edges in the adjacency matrix needs to be exactly the same as the order of the edges in your data, because FieldTrip assumes this order to be the same (and it has no way of checking or correcting for a possible mismatch).

So, long story short, with some fiddling around, what you want to achieve might be possible, but it needs some creativity on your end.

Best wishes,
Jan-Mathijs





On 27 Aug 2021, at 15:32, Julian Amacker via fieldtrip <fieldtrip at science.ru.nl<mailto:fieldtrip at science.ru.nl>> wrote:

Hi all,

I’m investigating methods for statistical analysis of intra-subject study. The data are PTE matrices (chan x chan)  for two conditions, effect and baseline.
Is there a function in fieldtrip to do a (spatial) cluster based permutation test on connectivity data? Or would it be possible to use the function t_freqstatistics.m but for connectivity matrices?

Many thanks and best regards,
Julian
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://urldefense.com/v3/__https://doi.org/10.1371/journal.pcbi.1002202__;!!HJOPV4FYYWzcc1jazlU!tfT_LCoZYiVD1NUCxB99ngsq2_9TFaCR5kW5cNTsmNSOJYWETsEnAOg769Yj823qqt3XYv3jJ3DCgwA$

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20210830/c4df3f03/attachment.htm>