Multi-Brain Computing: BCI Monitoring and Real-Time Decision Making

In this chapter we survey recent research on multi-brain applications. That is, applications in which synchronized brain activity of multiple users is measured and integrated in order to use their joint brain activity to make real-time decisions about communication with and control of devices in smart environments. Interestingly, we can go back to early brain-computer interface research of 1970 to see many ideas and sometimes implementations of synchronized multi-brain "computing". Usually they can be found in the artistic domain. In this decade (2010-20) we see growing attention in this research area, partly because of the availability of affordable electroencephalograhic (EEG) devices and partly because of the interest of human-computer interaction researchers in affective computing. This additional interest is now responsible for a focus on brain-computer interface (BCI) research that has changed from clinical applications to applications that are of interest to industry, specific groups of professionals or to the general population. Traditional BCI researchers are not always open to these developments, in which rather than focusing on Amyotrophic Lateral Sclerosis (ALS) patients, this new research focuses on entertainment, games, art and playful applications in the domestic and public domains. Among the many applications of multibrain computing: (1) Joint decision-making in environments requiring high accuracy and/or rapid reactions, or feedback; (2) joint/shared control and movement planning of vehicles or robots; (3) assess team performance, stress-aware task allocation, rearrange tasks; (4) characterization of group emotions, preferences, appreciations; (5) social interaction research (two or more people), (6) arts, entertainment, games. We discuss some examples from multibrain computing and focus on possible ways of joint decision making (or otherwise using the measured brain activity of multiple users). We will also emphasize the possibilities that are offered by the multimodal context, that is, considering brain-computer interfacing as one of the many possible modalities to obtain information about a user's or a group of users' affective states, preferences, and decisions. How to fuse information coming from different modalities and from different users need to be discussed. For that, we can learn from multimodal interaction research in human-computer interaction, including observations on sequential and parallel multimodality