Because there are some things you can't ethically test on humans, human medical research involves animal models. Such models are useful and important. There is a lot we wouldn't know—and a lot of people whose lives would be much worse—if it weren't for these animals.
That said, animals models are not perfect. Especially when it comes to relatively subjective problems like mental health. We test anti-depressants on mice. But how do you know whether it's working? After all, the mouse can't tell you that it's feeling better. And you can't really watch what the mouse does and see behaviors directly relatable to the human experience of depression, either. (Does the mouse feel more like going to his job and interacting with his friends today?)
At Scientific American Mind, Robin Henig explains the three commonly used tests that give scientists a glimpse into the mouse psyche. These are flawed proxies. Given the very real questions about how effective anti-depressant drugs actually are, it's worth putting some effort into developing better ways of monitoring their effectiveness in animals. But, for now, this is what we have to go on.
Forced swimming test. The rat or mouse is placed into a cylinder partially filled with water from which escape is difficult. The longer it swims, the more actively it is trying to escape; if it stops swimming, this cessation is interpreted as depressionlike behavior, a kind of animal fatalism.