The wearable system captures peripheral neural signals when internal speech articulators are volitionally and neurologically activated, during a user's internal articulation of words.
It doesn't capture the central signal, but peripheral, most likely from the nerves running around your face and your neck used to produce speech.