| We present an adaptive online brain-computer interface (BCI) based on natural reaching motor imagery with feedback of realistic animations of the intended reaching task.;We describe a unified approach to BCI design which addresses the complete system simultaneously to improve the overall user experience. In Chapter 2 we show how the proper selection of processing steps and combination of classifiers improves system accuracy. Chapter 3 discusses how realistic animations boost motor-imagery-related brain signals, and Chapter 4 describes classification rates in natural reaching tasks.;These findings come together in Chapter 5 where we describe our online system, which improves the user experience by eliminating training sessions, providing a very natural and straightforward mapping between user task and system feedback, and by using an adaptive classifier which can adjust to changes in brain signal over the time.;Research in this field has the potential to aid patients with motor disabilities, and it represents a novel way to study the brain. For users with severe motor disabilities such as ALS (Lou Gehrig's disease), a BCI may provide their only means of communicating with the world. Thus, reducing training time and improving the user's experience are important contributions to the BCI field. |
