Well, as you know, selective mutism is still poorly understood. Video self-modeling is effective for some children. Obviously not so much if a child is uncomfortable being videorecorded! When it works, the idea is that there are settings where speaking is comfortable, so take that language and use video production to change the perception of speaking in a new setting.

Of course, reducing anxiety is still the first goal.

...pronounced like the long vowel and first letter of the alphabet...