This seems relevant to LessWrong, both as an extreme example of how biases can hurt people and as a possible rationality technique. Depression is presumably at the outer end of some spectrum; to the extent that it's caused by cognitive mistakes, people in the middle of the spectrum should be able to benefit from undoing the same mistakes.
http://www.sciencedaily.com/releases/2011/11/111117202935.htm
Here's a link to the paper: http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=8436775