Background 
Biofeedback is the process of measuring an individual’s physiological states and feeding information about these changes back to them so that they can learn how to regulate their physiological activity to improve their wellbeing (Gilbert & Moss, 2003). In biofeedback training, participants are given insight into changes in their physiology such as heart rate, breathing or even brain activity by showing them visualisations (e.g., moving graphs of changes in heart rate) of these changes (Gilbert & Moss, 2003; Schwartz & Andrasik, 2017). These visualisations contribute to participants’ awareness of internal body signals, also known as interoceptive awareness, which is an important part of effective emotion regulation (Gross, 2002; Kever et al., 2015). In addition operant learning principles such as conditioning and reinforcement are used to teach participants how to effectively regulate their physiology. For example, a pleasant tone can indicate when someone’s physiology has reached an optimal level of activity (Gilbert & Moss, 2003; Hammond, 2005, 2007; Lehrer et al., 2000; Peper et al., 2009; Schwartz & Andrasik, 2017). 

While biofeedback has been shown to be effective as a treatment for anxiety (Goessl et al., 2017; Hammond, 2005; Richardson & Rothstein, 2008; Schoenberg & David, 2014; Schwartz & Andrasik, 2017; Tolin et al., 2020; Yucha & Montgomery, 2008) there are some practical issues that prevented a wide adaptation including specific hardware requirements and high costs. Furthermore, biofeedback training involves many trials which requires a huge time commitment and the trials are fairly repetitive making it difficult for participants to stay motivated, in particular when it comes to young participants (e.g. (Dedeepya et al., 2014; A. Parnandi & Gutierrez-Osuna, 2017). Another drawback of biofeedback training is that there is little understanding of the mechanisms of change; the majority of studies testing the effectiveness of biofeedback interventions focus mainly on outcome measures, such as symptom reductions (Weerdmeester, Van Rooij, Engels, & Granic, 2020; Wheat & Larkin, 2010). However, in order for treatments to be strengthened and tailored, it is crucial to know what factors contribute to successful change (and which ones do not) (Kazdin, 2014).The current project aimed to adress some of these barriers by exploring the potential of using biofeedback video games for anxiety regulation and to to identify possible mechanisms of change in biofeedback interventions for anxiety regulation. 

Our research 
In the past four years, most of our research has focused on assessing the potential of DEEP, a biofeedback-based virtual reality (VR) game. In this game, players move through an enchanting underwater world by using their own breath. This game was created by Owen Harris and Niki Smit (Monobanda PLAY) to provide players with a sanctuary where they can de-stress. Within this project, the game is currently being further developed and validated as a possible intervention for children with anxiety. DEEP stimulates players (through biofeedback) to use deep, diaphragmatic breathing. It is played by using a VR headset combined with a customized belt that measures the expansion of the diaphragm. The values of this sensor influence the game and are reflected in feedback in various ways. First of all, the player is informed about their breathing by way of their movement. Breathing in deeply gives the players an upward (when close to the ground) or forward force, and these forces are strengthened by deeply breathing out. The player is also given various forms of visual feedback. For example, a circle in the center of the screen shrinks and expands in accordance to the expansion of the player's diaphragm. In addition, elements in the environment (such as plants) change color or change in size/movement, mirroring the player's breathing. By giving this (bio)feedback to the player, they can become more aware of their breathing and are stimulated to breath in a deep and calm manner. 

So far we have conducted multiple studies including pilot studies, an implementation trial in special eduction, and most recently a Randomized Controlled Trial to assess whether DEEP could help alleviate anxiety. In addition we wrote an extensive review regarding biofeedback interventions for anxiety regulation. In this review we identified several psychological mechanisms of change that may be tied to adaptive outcomes of biofeedback interventions and we included a set of guidelines to for a new phase of biofeedback therapy, inspired by game design and the potential of wearable biosensors.