I reached out to the WSC team and will respond here when I receive a response. Thanks for using the site!
I would read the primary publications for the datasets, which may describe the general timeframe for different measurements. Some datasets will contain specific variables for "time between X (e.g., randomization) and Y (specific measurement)". The datasets may not wholly describe the sort of nuance you've hit upon.
The sleep studies in CHAT and BestAIR were done at different timepoints than the clinic visits that included the blood draws.
The CHAT/BestAIR baseline/screening sleep studies would often be conducted weeks before the blood draw at the baseline/randomization visit. The follow-up sleep studies may have been conducted close in time to the blood draws, though not always.
The CFS protocol was special in the way that measurements took place before and after the scheduled in-lab sleep study.
Thanks for checking out the site!
Thanks for using the site. The followup folder should contain home sleep test data from both 6-month and 12-month "final visits" (https://sleepdata.org/datasets/bestair/variables/final_visit). Subjects should only have one or the other (i.e., none with both 6-month and 12-month), or follow-up should be missing entirely.
The nonrandomized folder contains subjects who were screened with home sleep testing but were either 1) found to be ineligible or 2) opted not to continue to randomization. The baseline and followup folders contain data from randomized subjects.
It's referring to the signal metadata in the EDF headers for the home-based sleep studies. Sometimes sleep software developers do not output EDF headers to meet the EDF specifications.
I opened the HomePAP EDF files in EDFbrowser to check - https://www.teuniz.net/edfbrowser/
You could also use software like (Luna) to systematically output header information (which contains signal sampling rates) for a batch of EDFs - https://zzz.bwh.harvard.edu/luna/
The Pleth signal sampling rate appears to vary between the clinical sites that participated in HomePAP. I opened up a handful of files and found 25 Hz, 64 Hz, and 100 Hz.
Thanks for checking out the site.
nuMoM2b was a home sleep test that did not include EEG, so full sleep staging was not possible. The N2 indicators you see in the XML sleep staging represent estimated sleep time.
We stopped creating the "NSRR XML" files a few years ago. These type of annotation files were specifically created for some EDF-related tools that are now defunct, e.g., https://github.com/nsrr/edf-editor-translator
The Profusion and NSRR XML files contain the same information; the NSRR XML were derived from the Profusion XML. The NSRR XML formats and names things differently, particularly in regards to the sleep staging information.
Moving forward, we intend to use the .annot format as we've done in APPLES. Over time we will add these sort of .annot files to existing datasets like nuMoM2b.
Thanks for using the site!
The MNC data contributor has not deposited demographic data and has no plans to do so. This is a limitation of the dataset.