Electroencephalography (EEG) is the oldest and most accessible method for recording brain activity. Electrodes placed on the scalp detect the aggregate electrical signals produced by large populations of neurons firing together. First demonstrated in humans in 1924, EEG remains a cornerstone of clinical neurology for diagnosing epilepsy, sleep disorders, and brain injuries — and it has become the dominant sensing modality for non-invasive brain-computer interfaces.
EEG-based BCIs interpret patterns in scalp-recorded signals to infer a user's intent. Common paradigms include steady-state visually evoked potentials (SSVEP), P300 event-related potentials, and motor imagery — where imagining a hand movement produces detectable changes in brain rhythms. Companies like Emotiv, Muse (InteraXon), and OpenBCI sell consumer and research-grade EEG headsets that enable applications ranging from meditation feedback to hands-free computer control.
The fundamental limitation of EEG is its low spatial resolution and weak signal strength. The skull and scalp tissue act as filters, blurring and attenuating the neural signals before they reach the electrodes. This limits the information transfer rate of EEG-based BCIs to roughly 10-25 bits per minute in practical use — far below what invasive interfaces achieve. Advances in dry electrode technology, machine learning-based signal processing, and high-density electrode arrays are gradually improving EEG performance, keeping it relevant for applications where non-invasiveness is paramount. For deeper coverage, see BCIIntel.