RE: segmentation of functional data

Hello, 

once I denoise the data by means of the masks stemming from segmenting the functional data, how can I assure that this procedure works well?

What we did was to extract time series from a set of ROIs and estimate the correlation matrix. We did that from the denoised images with aCompCorr. However, the correlation between the two was only 0.3 (though significant). 

Shall we trust this procedure anyways? 

Any tips?

Thanks so much, 

Athena