This present book covers numerous examples of advanced machine-learning and signal processing algorithms to robustly decode EEG signals, despite their low spatial resolution, their noisy and nonstationary nature. These algorithms are based on a number of advanced techniques including optimal spatial filtering, tangent-space mapping, neural networks and deep learning, transfer learning, parametric modeling, supervised connectivity analysis, supervised and unsupervised adaptation, and incorporating signal structures, among many others. Importantly, this book goes beyond the EEG-decoding challenge and discusses the importance of using signal processing and machine-learning methods to model and update the user's sates and skills over time. These user's models could help design a better EEG decoder (features and classifier) that not only leads to a good discrimination of BMI commands but also facilitates user learning.