Thanks for these comments Rachel. Great idea to use information on mother-infant contact - perhaps that is something that is very much shaped by arbitrary aspects of hospital policy so you could find a natural experiment. One challenge might be that the relevant etiologic period for human infants might be much longer than for rat babies.
There is some evidence on maternal depression (maybe even prenatal?) and children's outcomes, but it's tough to understand whether that's mediated by maternal-child interactions or confounding by whatever increases the mother's risk of depression. Still, an interesting approach to get some leverage on the problem.