Details
-
New Feature
-
Status: Closed
-
Minor
-
Resolution: Fixed
-
1.0.1
-
1.5
Description
It would be nice to display the results of ICA as spatial maps, i.e.
displaying the weights of a certain component over the brain surface
I imagine it could adapt depending on the "source" TimeSeries:
- TimeSeriesRegion dtypes: - graph displays like the 2D Connectivity visualizers (the size/color of the nodes is given by the weights).
- when selecting a node the component time-series (scaled by the corresponding weight) is displayed.
- OR for a given component, display in a topology map its value for every region (applying some interpolation and smoothing) – see attached figure visualizer_ICA
- TimeSeriesSurface dtypes: - my guess is that in this case displaying the component loadings is not very relevant.
- displaying the surface (3D view as in Local Connectivity/Surface Stimulus) for a few components.
Issues:
..........
Selecting # first dominant components based on the non-normalized mixing matrix (L_
{infty}-norm?).
Need to add references and develop more