Hi!

I just copy the thread on github here, so it's easy to search. Thanks for your help @Francois!

**Question:**

Is there a way to cluster sensors in TF maps? I see there's the option to cluster the time series, but not for TF maps. Should I need to do it for the time series and then compute again the TF?

Similar, is there a way to group in time bands for time series? I see there's an option to group in time and frequency bands for TF maps, but not for time series.

**Response:**

Is there a way to cluster sensors in TF maps?

You can use the process `Extract values`

to select some sensors and average their TF response together:

This must be done separately for each cluster.

Should I need to do it for the time series and then compute again the TF?

This would produce something very different.

Similarly to averaging trials, you should probably average the TF power instead of compute the TF of the average across sensors.

is there a way to group in time bands for time series?

No. The "data" or "matrix" file structures do not have any information for grouping by time bands.

https://neuroimage.usc.edu/brainstorm/Tutorials/Epoching#On_the_hard_drive

You can use the process `Average time`

to compute the average over a time window.