On the production of knowledge in sleep laboratories and via self-tracking