Back to the main page.
Bug 2728 - addition to cluster_permutation_freq tutorial
| Status | CLOSED FIXED |
| Reported | 2014-10-07 21:56:00 +0200 |
| Modified | 2017-01-17 11:30:38 +0100 |
| Product: | FieldTrip |
| Component: | documentation |
| Version: | unspecified |
| Hardware: | All |
| Operating System: | All |
| Importance: | P5 minor |
| Assigned to: | Jan-Mathijs Schoffelen |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Jens Klinzing - 2014-10-07 21:56:42 +0200
In this tutorial in the code examples at #plotting_the_results and #plotting_the_results1 the cases should be considered that a) one wants to plot negative and positive clusters (this is only the case in the first example), and b) that there are either no negative or no positive clusters. I would suggest to enclose the relevant parts of the code with an if-statement: if ~isempty(stat.posclusters) pos_cluster_pvals = [stat.posclusters(:).prob]; pos_signif_clust = find(pos_cluster_pvals < stat.cfg.alpha); pos = ismember(stat.posclusterslabelmat, pos_signif_clust); else pos = false(size(stat.prob)); end I can add that myself, I just wanted to ask before.
Jan-Mathijs Schoffelen - 2016-12-23 11:02:03 +0100
Hi Jens, I am in the process of cleaning up old bugs which have been open for a long time. In the meantime it could be that the relevant code has changed anyhow... This bug here inspired me to look in the tutorial, where I changed the wording such (hopefully) that nowhere is mentioned anymore 'significant clusters'. Yet this was not the purpose of the original bug report, but I think that your suggested code changes pertain to ft_clusterplot. It seems that the code has developed since 2 years, so I am not sure whether your original question/concern still applies. I suggest for now to close this one, unless you strongly feel about reopening it: in which case feel free to do so.