Grant: Discovery Grants Program - Individual

Brain-computer interface (BCI) technologies have been traditionally defined as devices that enable movement-free communication through the direct measurement and translation of brain activity that is intentionally generated by the user for the purpose of controlling an external device. The target users of such ""active"" BCIs are generally individuals with severe motor impairments who are unable to operate more conventional interfaces (e.g., keyboards). Although this remains the focus of most BCI research, the potential of BCIs for healthy users has also been recognized and is receiving increased attention. The practical definition of BCIs is expanding to include not just those devices using brain activity to replace natural central nervous system (CNS) output in users with disability, but also those that enhance CNS output in healthy users. Rather than exploiting intentionally modulated brain activity, these so-called ""passive BCIs"" derive their outputs from arbitrary brain activity arising without the purpose of voluntary control, and are used instead for detecting the mental state (e.g., cognitive, affective) of the user. Passive BCIs have potential in many applications to enrich human-machine interaction by providing implicit information on the user's state and adapting the environment accordingly. One application domain that has received particular attention is the monitoring of mental workload, particularly in safety-critical occupations like pilots, air-traffic controllers, and other industrial operators. The ultimate goal is to derive a continuous, objective estimation of the cognitive strain on the individual from physiological signal variables, so that appropriate action can be taken to reduce the potential for error during periods of extreme demand or overload. Mental workload detection also has potential value in other domains, including gaming, adaptive training, and user interface design applications, to enhance and personalize user experience. Previous studies support the feasibility of a passive BCI for automatic mental workload detection. However, to date only very basic functionality has been explored - specifically, the automatic detection of brain activity associated with different levels of difficulty of a given task under controlled conditions. While this is a critical first step, there is still much work to be done before such systems are suitable for widespread practical application, and much opportunity remains for enhancing reliability and functionality. In this research program, we will build upon the current state-of-the-art in passive BCIs for automatic mental workload detection. Specifically, we will work toward developing a system that is both task- and emotion-aware. The enhanced capability will result in improved user-state prediction, which will in turn allow for enhanced human-machine interaction in a variety of applications.