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.