ICA with ROI analysis

I want to view only specific ROIs in my ICA plot ; I have computed ICA using nilearn package, but it is giving me all the activated regions. I want to find out if some specific regions are activated in a subject and thus I want to use 2 of the MSDL atlas’s regions. Can someone guide me about how to mask other unwanted regions from the ICA plot ?

You’ll need to basically zero signal not covered by the ROIs you want to extract, before plotting. I’m not sure that is something you really want to do.

Alternatively, you can have your components plotted as they come from ICA but draw the contours of the ROIs you are interested in. You can maybe start looking at this example: https://nilearn.github.io/auto_examples/04_manipulating_images/plot_extract_rois_statistical_maps.html

Is there a way to automatically label the regions found in ICA ?

What do you mean by automatically label the regions found in ICA ?

I mean that is there a tool that can separate out the regions highlighted in ICA plot saying which one is PCC, IPS etc. ?