A BCI captures either invasive signals (electrodes implanted at or near the brain's surface) or non-invasive signals (external sensors reading brain activity, or residual muscle activity in cases of paralysis) and feeds them into a decoder trained to map that signal to an intended action, word, or letter. Because these signal patterns vary significantly between individuals, practical BCI systems are typically trained or fine-tuned per user rather than shipped as one universal model. Current applications span mobility control, text input, and speech restoration for people with paralysis, ALS, or locked-in syndrome; a BCI decoding attempted speech and feeding it into a voice-synthesis system is a distinct combination from a BCI decoding movement intent for cursor or wheelchair control.