Device Interaction with Non-Invasive BCI

Abstract - This review focuses on Non-invasive Brain–Computer Interfaces (BCIs), which are systems designed to establish a direct link between the brain and external devices without the need for physical movement. These EEG-based BCIs are applicable in assistive and interactive contexts. A significant emphasis is on BCI-based spellers, which facilitate communication through brain signals. The review examines three main types: P300, Motor Imagery (MI), and Steady-State Visual Evoked Potentials (SSVEP). P300 spellers utilize event-related potentials for high precision; MI spellers depend on imagined movements but demand extensive training; and SSVEP spellers provide rapid performance using flickering visual stimuli. This review details various EEG signal processing techniques, BCI system architecture, and contemporary classification methods for EEG-based BCI. Key Words: Electroencephalography (EEG), Brain-Computer Interface, Speller, P300, Motor Imagery (MI), Steady State Visual Evoked Potential (SSVEP)