Dear Alfonso,

I would like to ask an additional question on this regard.

I am in a similar situation, where my aim is to create an overall network connectivity score for each of the 7 networks in the Yeo atlas.

I was wondering if it is sensible to compute the average, as you suggested to Reka in the previous message, or transforming all Z scores in their absolute values first (therefore getting rid of the directionality of the connectivity), and then computing the average afterwards? 
In my case, I am mostly interested on the magnitude, or connectivity strength, rather then on the directionality.

Any suggestions on this regard would be most helpful.

Thank you very much!

Michela

Originally posted by Alfonso Nieto-Castanon:

Dear Reka,

Yes, computing the average of those 21 Fisher-transformed connectivity values is a perfectly reasonable/meaningful marker of within-network connectivity (see for example conn_withinbetweenROItest.m for a similar computation/analysis)

Best

Alfonso

Originally posted by Réka Borbás:

Dear Alfonso and conn users,

Using a ROI-to-ROI approach, is it meaningful to calculate an overall network connectivity score based on the individual ROI-to-ROI fiher-z scores? 

To be more specific: 
I have a network of 7 ROIs (defined by me). For each subject I have 21 fisher-z transformed correlation scores representing all possible connections within the network. If I calculate the average of these 21 scores per subject, can it be a meaningful marker of overall network connectivity? 

Thank you in advance,
Réka