search

0todd0000 / spm1dmatlab

One-Dimensional Statistical Parametric Mapping in Matlab.
GNU General Public License v3.0
27 stars 13 forks source link

GLM with 1D covariate? #193

Closed danitrowell closed 5 months ago

danitrowell commented 9 months ago

Hi Todd,

I hope you are well!

We are investigating the effects of an intervention (pre v post) on 1D overground running biomechanics between two groups (intervention v control). We would like to control for differences in overground running speed between baseline and follow up testing.

Are any of the SPM1D packages suitable for this design? I trialled using a GLM but I don't believe it's possible to use a 1D variable as a covariate. Is my understanding correct? I.e., determine intervention-control difference on 1D biomechanics (e.g., ankle angle at follow-up) while accounting for two covariates: 1D biomechanics at baseline and the difference in running speed (post-pre speed).

Kind regards, Dani

0todd0000 commented 9 months ago

Correct. A 1D covariate would require a type of cross-correlation analysis that is called "connectivity analysis" in the SPM literature. This may be supported in a future version of spm1d but it is currently not a development priority. At the moment it is only possible to use 0D scalars as covariates via GLM.