help
help > RE: Association with the connection strength
Apr 7, 2020 08:04 AM | Andrew Zalesky
RE: Association with the connection strength
Hi Nabin,
Option 1: you could extract the connectivity values for the significant connections and then test for a correlation with you behavioral measures post hoc. Instructions are provided in the manual to extract connectivity values from a significant subnetwork.
Option 2: You could include your behavioral measure as a column in the design matrix. You could then look at the interaction between group and behavior. So your design matrix would have one column for group, another for behavior and a third (interaction) formed by multiplying the group and behavior columns.
I hope that helps.
Andrew
Originally posted by Nabin Koirala:
Option 1: you could extract the connectivity values for the significant connections and then test for a correlation with you behavioral measures post hoc. Instructions are provided in the manual to extract connectivity values from a significant subnetwork.
Option 2: You could include your behavioral measure as a column in the design matrix. You could then look at the interaction between group and behavior. So your design matrix would have one column for group, another for behavior and a third (interaction) formed by multiplying the group and behavior columns.
I hope that helps.
Andrew
Originally posted by Nabin Koirala:
Dear Andrew,
I apologize in advance as it might sound odd what I am trying to do but wanted to know if it is possible from NBS, as the answer would be very interesting.
I have 2 groups and I did the NBS using the Diffusion data output (t-test) and found that there are several connections significant different between the group. Now i wanted to know if one or more of these connection strength (t-stat value) is associated with some Behavioral measures from each of the group. I have the values for each subjects for both groups but could not think about the design matrix that would do this. Would making the values as covariates and see the association to them give me what I am looking for, if so how do I get different association for each group ? Your suggestion in the right direction would be really helpful.
Thanks a lot in advance.
Best regards,
Nabin
I apologize in advance as it might sound odd what I am trying to do but wanted to know if it is possible from NBS, as the answer would be very interesting.
I have 2 groups and I did the NBS using the Diffusion data output (t-test) and found that there are several connections significant different between the group. Now i wanted to know if one or more of these connection strength (t-stat value) is associated with some Behavioral measures from each of the group. I have the values for each subjects for both groups but could not think about the design matrix that would do this. Would making the values as covariates and see the association to them give me what I am looking for, if so how do I get different association for each group ? Your suggestion in the right direction would be really helpful.
Thanks a lot in advance.
Best regards,
Nabin
Threaded View
| Title | Author | Date |
|---|---|---|
| Nabin Koirala | Apr 6, 2020 | |
| Andrew Zalesky | Apr 7, 2020 | |