Hi Guys,
We're trying to sync respiratory events, O2 desats and HR data from the SHHS and having a similar issue, so I was wondering if you could help with a couple of questions.
Firstly, I saw in another forum thread (https://sleepdata.org/forum/ahi-definition-in-shhs/1#comment-677) that ahi_ah03a was created at a later date with hypopneas only being counted at a 3% desaturation threshold. How did the software calculate this if the O2 desaturation annotations are incorrect? It seems from your comments that there's no way to get a list of hypopneas with a 3% or 4% desat (only the overall AHI for each)?
Secondly, I see above that "SpO2 desaturation annotations and linking with respiratory events were affected". Do you know if they were shifted in a linear fashion? i.e. SpO2 desats are all 30s later. Or were they randomly affected/unlinked to the events? (if shifted in a linear fashion, would be fairly easy for us to correct the times).
Finally, although the O2 desat in the XML files may be incorrect, are the respiratory events in the XML annotations correct? Or more specifically, do their start times match up correctly with the EDF files? Or were the respiratory event times also affected by the software changes over time?
Essentially we want to see if respiratory events (our definition of RE = any apneas + hypopneas resulting in a 4% desaturation or arousal) result in HR changes. If I am understanding the known issues correctly, we would have to: 1) create an O2 desaturation algorithm from the EDF data to accurately link to respiratory events, thus allowing us to remove "false" hypopneas (i.e. <4% desat), 2) then use our program to see if respiratory events result in HR changes. However, if the respiratory event times are also desynchronized from the EDF files (and thus HR) this would make the analysis incorrect and would require us to re-score the EDFs.
Any insight into the above would be greatly appreciated! Thanks for your time.
Joel (Research Assistant working for Dr Sankari)