[Eeglablist] Clustering Problmes - update

[wibral at mpih-frankfurt.mpg.de]

Dear Listusers,

I have meanwhile tried quite a few combinations of IC features to
determine the distances used in clustering and also toyed with the
dimensionality of the feature vectors, but there seems to be a problem
when clustering > 18 components, independent of what I feed the
algorithm.
I hope this helps to track down the problem. Clustering with up to 18
Clusters works well, although the numbers of components in the clusters
vary widely (of course).

There is another, possibly related error: When I try to plot the scalp
maps of the components of a cluster, I get the following error message:

Drawing components of cluster (all at once)...
??? Attempted to access grid_pol(1,2); index out of bounds because
numel(grid_pol)=1.

Error in ==> std_topoplot at 148
                grid_pol = sign(grid_pol(1,2));

Error in ==> pop_clustedit at 484
                    eval(['STUDY = std_' plotting_option '(STUDY,ALLEEG,
''clusters'','  num2str(cls(clus-1)) ', ''mode'', ''comps'' );'  ]);

??? Error using ==> pop_clustedit('plotcomptopo',gcf); 
Attempted to access grid_pol(1,2); index out of bounds because
numel(grid_pol)=1.

??? Error using ==> waitfor
Error while evaluating uicontrol Callback


Best Regards,
Michael Wibral