Brain-computer interfaces (BCIs) are devices that process a user’s brain signals to allow direct communication and interaction with the environment. BCIs bypass the normal neuromuscular output pathways and rely on digital signal processing and machine learning to translate brain signals to action. Brain signals are recorded either noninvasively from electrodes placed on the scalp [electroencephalogram (EEG)] or invasively from electrodes placed on the surface of or inside the brain. BCIs based on these recording techniques have allowed healthy and disabled individuals to control a variety of devices. Learn about the challenges and proposed solutions for noninvasive brain-computer interfacing through IN THE SPOTLIGHT column of the IEEE Signal Processing Magazine (July 2010).