I've done a lot of writing and analysis in health sciences. My very strong intuition is that there is a study flaw here of one form or another.

What this seems to say, to me, is that people who choose to eat healthy also tend to be less depressed than those who choose to not eat healthy.

Population bias may be a problem here - that eating unhealthy may not necessarily cause depression, but rather those who are eating unhealthy are also likely to be suffering from depression.

There is limited data on diet and its impact on health in general - except in obvious cases of frank deficiency.

This could be like saying "People who go jogging daily tend to be happier in general." You could say that jogging then causes happiness. Or you could say that people who tend to jog are demographically likely to also do happy inducing things.