PhD Defense: Designing Brain-Computer Interfaces for Intelligent Information Delivery Systems
While traditional systems use static designs that are meant to reach as many people as possible, a user's moment-to-moment state can impact interaction, either positively or negatively. In this thesis, I show that brain sensing can be used as passive input to intelligent information delivery systems. Using functional near-infrared spectroscopy (fNIRS) as a lightweight brain sensor, I present work to design brain-computer interfaces (BCIs) that analyze fNIRS data and classify user state in real time. These systems react to user state by modifying the flow of information and measurably improving user performance. I describe a brain-driven recommendation system that changes which information is prioritized to the user, an interruption management system that users brain data to determine opportune moments of interruption, and a study that demonstrates the brain's sensitivity to visual design. Finally, I discuss design strategies for building robust online systems that adapt to physiological input. I suggest that someday our computers may have the capability of being personally attentive to us - optimizing when information is delivered, which information is prioritized, and how information is presented.