Silent speech interface
A system that lets a user communicate by speaking silently, mouthing or attempting speech without phonation. Historically the term spans non-neural sensors: surface EMG of the face and neck, ultrasound or electromagnetic articulography of the tongue, and lip video. In the BCI context it denotes cortical decoding of silent or attempted speech.
The unifying goal is speech output without an audible voice, for privacy, for noisy or quiet environments, or, most importantly here, for users who cannot phonate at all.
Neural silent speech (attempted speech decoded from cortex in an anarthric user) and peripheral silent speech (facial EMG in an able-bodied speaker) address related goals with entirely different signals and populations; their accuracy figures are not comparable.