Your query on Sleep Quality measuring device implies a domain within which to measure quality.....even your qualifier of polysomnography as a 'gold standard' has it many limitations. Sleep is multi-dimensional process and the rich research literature shows that to reduce it to a single domain will continue to reflect this limitation. Even using EEG as reflected in the widely used R&K polysomnograpgy
Franklin Lue sure, sleep is multidimensional. Still, objective measurements that include accurate depictions of different times connected to sleeping (time-to-sleep-onset, time in bed, total sleep time, time spent in various sleep zones) gets us closer to measuring and quantifying the REAL sleep. The closer the better. My question is: Which currently available device comes closest.
I know of a couple of non-contact sensors that use low power radar waves (ResMed S+) and bed sensors (Beautyrest) that all claim to be accurate but fail in actual comparison to PSG. Also, I have come into contact with Danish scientist @Kaare Mikkelsen who is researching different types of EEG systems that prove to yield as accurate a depiction of sleep as the classical EEG used in PSG in a sleep laboratory.
I've been out of the sleep research field for quite a few years now so am not up date on the latest literature. But I am confident, that if you even scan summaries of the sleep literature, you will find very poor predictability of sleep quality based on any single device. If what you are referring to by "Real sleep' as motion detection then i would focus on sensor reliability. Even the R&K standard scoring PSG have limited predictability for patient complaints of sleep quality; meaning sleep complaints.
Thank you so much, Franklin Lue , much has changed since we last spoke. Check out this new research which compares a non-contact sleep monitor to polysomnography: Article Validation of sleep stage classification using non-contact r...
Even though, you have been sceptical towards even the validity of PSG, would you think that measuring sleep without interrupting the subject would yield better results?