mrueschman

The columns appear to be there in the 'followup' CSV when I look in Excel or in a text editor. E.g.

What are you using to open the file?

Hey Pengom,

Thanks for checking out the site. The PAP therapy variables (e.g. https://sleepdata.org/datasets/heartbeat/variables?folder=Measurements%2FPAP+Therapy) only exist in the heartbeat-followup-dataset-0.4.0.csv dataset.

I don't think there was an explicit question about heart failure. The closest would probably be dxheartattack.

From what I can tell in the data processing code, "ACE Inhibitors or ARBs" were included in the aceinhibitor variable.

I don't think the sleep scoring software output ODI values. There are some variables about number of desaturations of different magnitudes, but I'm not sure it's possible to compute an ODI3/ODI4.

Time spent with SpO2<90%: satlt90p

Best,

Mike

Hello,

Are you looking to find datasets that meet these criteria? Many of the datasets here have HR signals. Look for the "montage and sampling rate information" pages in the dataset documentation to see available signals. The MESA dataset also has R-point annotation files available. Most of the datasets have manual 30-second epoch staging available in the XML annotation files.

Mike

Hi - have you tried rewriting one of the EDF headers to have a valid date (e.g. 01.01.85 or 01.01.01)? I have seen the same error about 00.00.00 when using EDFbrowser.

Here's a tool that you could try using to rewrite the header in one of your files and then test again in EDFbrowser/Pyedflib - https://prerau.bwh.harvard.edu/edf-de-identification-tool/

Good luck!

Thanks for checking out the site. None of our currently posted datasets contain audio recordings (there might be challenges with this because all our data are deidentified). There are "Snore" signals in some datasets which were collected either via nasal cannula or other sensor that detects vibratory snore.

Hi -- this question came up before -- https://sleepdata.org/forum/mesa-sleep-position

• 0 - Right
• 1 - Back
• 2 - Left
• 3 - Prone
• 4 - Upright

I'll make a note to add it somewhere in the MESA documentation as well. Thanks for checking out the site!

The technicians worked separately - each sleep recording was staged by a single technician. The group of technicians had routine meetings and quality exercises to ensure consistency over time across scorers.

You can read a bit more about the reliability/quality/consistency procedures used here: https://sleepdata.org/datasets/shhs/pages/07-reliability-shhs2.md

Thanks for being an NSRR user and for asking this question!

I see what you mean - it looks like all the EDFs received from one of the CHAT sites/beds have this quirk. Unfortunately, many PSG manufacturers do not follow the EDF specifications to the letter. Other tools like Polyman and Luna do not seem to have issue opening these files.

Based on my reading of the EDF FAQ Q8 is that you could theoretically put dummy values in the physical min/max (i.e. rewrite this part of the EDF header for affected files) to overcome this issue.

Hi sleepheart,

Thanks for being a user here. Unfortunately we (the NSRR) are only able to offer MESA data that is related to the sleep exam, which was an ancillary study to the main Exam 5. See the note here: https://sleepdata.org/datasets/mesa

MESA Sleep data are made available in collaboration with the MESA Coordinating Center. We have made polysomnography and actigraphy raw data available for download, along with key sets of MESA Sleep covariates. If you wish to access data from other MESA exams, please visit the BioLINCC MESA Collection or contact the MESA Coordinating Center.

Those resources will have their own data request processes.

Hi - thanks. Can you share more about the error? I refreshed the folder on the NSRR system (generates md5 checksums and caches filenames) and now the command appears to work:

There were some recent compatibility issues in the NSRR gem with Ruby 3.0.0 - those got patched in version 6.0.0.

Please give it a try yourself to see if the issue is resolved. Thanks for being an NSRR user and raising this issue.