The researchers used a system called MobiHealth Mobile. This is what it looks like:
It's a lot less bulky than some biofeedback devices I've seen. Those little sensors probably go on your fingers.
Okay, so you have data about what the player's mood is. What can you do with that data? Well, you can tell when they're getting frustrated, and send them to meditate until they physically calm down.
When high undesired emotional and/or physiological reactions (e.g. anger feelings, impulsiveness, non-relaxed reactions, frustration, quick and unplanned responses) are detected by the video game, the game immediately directs the avatar to a relaxed area with the goal to calm down. During the whole game session, higher undesired emotional and/or physiological reactions are coupled with greater difficulty to reach the end goals of the video game (e.g. while diving the fishes are more difficult to catch, more obstacles appear in the mini-games). More relaxed and self-controlled reactions are positively reinforced by the game, making the situations easier to handle and the end goals easier to reach.In their game PlayMancer, the frustrating minigame is trying to collect things underwater while maintaining their oxygen level. In the calming game, more stars appear based on how relaxed you are.
Now, what if there were a game that undermined these goals? You could make a boss get tougher and tougher based on how frustrated (or how calm) the player was. Or a game where you don't die until your palms are sweating with anger. Or a dating sim that only gives you the suave lines if you're really calm. THINK OF THE POSSIBILITIES. Think of how immersive this would be combined with the Oculus Rift.
Here's a short gameplay video of Playmancer, which looks like some kind of a college senior project, but the impressive part about this game is the inputs, which unfortunately you can't see.
Source: "Video games as a complementary therapy tool in mental disorders: PlayMancer, a European multicentre study" by Fernando Fernandez-Aranda et al.