help > Quadratic model
Showing 1-5 of 5 posts
Aug 11, 2020 04:08 AM | Hugo Morandini
Quadratic model
Dear Alfonso,
I would like to calculate the non-linear effect of age on the connectivity of some ROI pairs, however, I am unsure about the way to do this in CONN.
I was thinking about creating a 2nd level covariate 'AgeĀ²' and to use it as a regressor of interest. Would that be a good way to proceed?
Your advice would be appreciated.
Thank you.
Kind regards,
Hugo
I would like to calculate the non-linear effect of age on the connectivity of some ROI pairs, however, I am unsure about the way to do this in CONN.
I was thinking about creating a 2nd level covariate 'AgeĀ²' and to use it as a regressor of interest. Would that be a good way to proceed?
Your advice would be appreciated.
Thank you.
Kind regards,
Hugo
Aug 15, 2020 03:08 PM | Alfonso Nieto-Castanon - Boston University
RE: Quadratic model
Dear Hugo,
Yes, that is a perfectly reasonable approach. You may for example create an age2 covariate (e.g. and enter in its values field "age.^2" to have CONN automatically fill-in those values) and then define a second-level analysis that includes constant, age, and age2 effects to evaluate possible non-linearities (and the same approach works for any base characterizing a family of temporal trajectories, not just polynomials).
Best
Alfonso
Originally posted by Hugo Morandini:
Yes, that is a perfectly reasonable approach. You may for example create an age2 covariate (e.g. and enter in its values field "age.^2" to have CONN automatically fill-in those values) and then define a second-level analysis that includes constant, age, and age2 effects to evaluate possible non-linearities (and the same approach works for any base characterizing a family of temporal trajectories, not just polynomials).
Best
Alfonso
Originally posted by Hugo Morandini:
Dear Alfonso,
I would like to calculate the non-linear effect of age on the connectivity of some ROI pairs, however, I am unsure about the way to do this in CONN.
I was thinking about creating a 2nd level covariate 'AgeĀ²' and to use it as a regressor of interest. Would that be a good way to proceed?
Your advice would be appreciated.
Thank you.
Kind regards,
Hugo
I would like to calculate the non-linear effect of age on the connectivity of some ROI pairs, however, I am unsure about the way to do this in CONN.
I was thinking about creating a 2nd level covariate 'AgeĀ²' and to use it as a regressor of interest. Would that be a good way to proceed?
Your advice would be appreciated.
Thank you.
Kind regards,
Hugo
Sep 14, 2020 04:09 AM | Hugo Morandini
RE: Quadratic model
Thank you very much Alfonso.
Just regarding the definition of the contrast, to look at the non-linear effect of age controlling for gender, would it be:
constant, age, age2, gender: [0 0 1 0] ?
Kind regards,
Hugo
Just regarding the definition of the contrast, to look at the non-linear effect of age controlling for gender, would it be:
constant, age, age2, gender: [0 0 1 0] ?
Kind regards,
Hugo
Sep 29, 2020 12:09 AM | Alfonso Nieto-Castanon - Boston University
RE: Quadratic model
Yes, exactly
Best
Alfonso
Originally posted by Hugo Morandini:
Best
Alfonso
Originally posted by Hugo Morandini:
Thank you very much Alfonso.
Just regarding the definition of the contrast, to look at the non-linear effect of age controlling for gender, would it be:
constant, age, age2, gender: [0 0 1 0] ?
Kind regards,
Hugo
Just regarding the definition of the contrast, to look at the non-linear effect of age controlling for gender, would it be:
constant, age, age2, gender: [0 0 1 0] ?
Kind regards,
Hugo
Sep 29, 2020 03:09 AM | Hugo Morandini
RE: Quadratic model
Thank you very much Alfonso.
Kind regards,
Hugo
Kind regards,
Hugo