Why is sex handled as two different variables in your model?

As for the cognitive composite measure, it might be better to convert the scores of the 5 variables to a common metric (z-score) and then derive the mean of those z-score converted variables for your executive composite.  

Jeff