Originally posted April 2018

MIT’s AlterEgo headset detects subvocal signals, processes signals through a neural network and translates the signals to words, then sends the audio via the bone-conduction part of the headset.  Good thing this wasn’t around when I was an angsty teen – constantly muttering my true opinions under my breath.  Whoops, I mean ‘praising my parents and teachers.’

TechRadar.com