Notes.txt

For this model many of the steps will work in conjunction with Dr, Chris Habeck's 
package. But some of the steps need to be modified or replaced when working with 
moderated-mediation models. In particular the model fitting which is perfored 
using the code: minAIC_fit_f.m which is called by the 
routine behav_fit_bootstrap_module_f.m at line #257. The calculation of 
the parameter estimates needs to be replaced in some models with a 
fminsearch operation at line #54.

Looking at Chris's code the permutation testing can take place on the cluster 
fairly easily because the key image is the variance image. And this image is 
calculated via summation over each permutation. Therefore the jobs can be 
split then put back together no problem.

04/5/2013
I need to modify subfnProcessModelFit so that it calculates the conditional 
effects at multiple values of the probe. This means I need to get rid of the 
if clause for checking the values of the interaction effect.