I like this idea, as well as the irony of the original study being not blind. ;)
Another possibility might be to design the functionality into the mask itself. When it detects the user is asleep, it could open up to let in light (or not) based on which cohort the user is in.
The detection could be done by having sensors in the mask itself, or it could just pair with the user's existing fitness/sleep tracker of choice via Bluetooth.