NSRR staff
Boston, MA
0000-0002-0506-8368
Top Topics
Recent Topics
I believe the STAGES signal names are not harmonized across sites, so you will find different signal labels in the EDF headers from site to site.
I suggest checking out the documentation provided in this folder as you go through the EDFs from different sites - https://sleepdata.org/datasets/stages/files/original/Site%20Specific%20PSG%20Notes%20and%20Information
ECG/EKG data collection may have differed between sites.
Hey DibishaTP - thanks for bringing your questions to the forum. Unfortunately, we didn't receive much documentation about the actigraphy device and output from STAGES. There was a brief discussion about it on the forum previously - https://sleepdata.org/forum/stanford-dataset-documentation/
Further, I don't think you'll be able to overlap the STAGES polysomnography and actigraphy data. The actigraphy data were provided to the NSRR as date-shifted files (random shifting amounts unknown to us), however the EDFs had all date references stripped.
Thanks Jeremy/Chao - good questions. In addition to the overall, combined AHI variables, most datasets will have obstructive apnea and central apnea indices alone, e.g. in SHHS --
If you look at the Calculations for these different indices you'll see that each event type has 4 components, which are individual event counts in REM/Supine, REM/Non-Supine, NREM/Supine, and NREM/Non-Supine.
Given that, you could recompute new AHI, AHIc, or other indices based on your final criteria of interest.
What you see on the "13-hrv-analysis" page is what we received from the data creators. We don't have any specifics about configuration settings.
I believe the SHHS data were run on an older version of the PhysioNET HRV toolkit - the original link no longer works but here's an archived copy: https://web.archive.org/web/20170615014052/https://physionet.org/tutorials/hrv-toolkit/
The latest tool, PCST (https://physionet.org/content/pcst/1.0.0/), has this note: "It [PCST] was shown to be equivalent to the PhysioNet C HRV Toolkit".
We generally remove calendar dates from all our files since they are considered identifiers. This does pose difficulties in aligning some data, such as the MESA PSG and actigraphy files.
There is a section on this page that describes the overlap between MESA PSG and actigraphy data - https://www.sleepdata.org/datasets/mesa/pages/actigraphy-introduction.md
I believe you could use that supplemental file, along with the HRV **seconds **column (https://www.sleepdata.org/datasets/mesa/pages/hrv-analysis.md), to align these data sources (when they overlapped, which was not the case for some subjects).
Thanks for the suggestion - I agree raw accelerometry data will be helpful to researchers. Unfortunately, the device used in MESA and HCHS - the Philips Actiwatch Spectrum - does not output raw accelerometry data. It only outputs activity/light data into binned epochs, e.g. 30 seconds. For future datasets with accelerometry data we plan to make those data available for download.
Thanks for checking out the site. In SHHS, HRV analysis was only performed during sleep time. From https://sleepdata.org/datasets/shhs/pages/13-hrv-analysis.md:
--
summary (Entire night) - Analysis is performed from sleep onset to sleep termination. This analysis requires the following information: the time of occurrence of each R-wave, and the sleep onset and termination times. These times were extracted from the file containing the sleep stage annotations as the first and last 30 sec episode of stage 1, 2, 3 or 5 (REM), respectively. RR intervals larger than 2.5 seconds were excluded from the analysis.
Thanks for your interest in the site. From https://sleepdata.org/about/
The NSRR Data Access Committee reviews incoming data requests and responds to them within 1-2 weeks of submission.
Thanks - I saw your question. I referred it to another team member to comment first.
I'm very confused about how to use the dataset. Can you give me some advice about how to use it correctlly?
Note the "Known Issues" for SHHS, if you haven't already: https://sleepdata.org/datasets/shhs/pages/05-polysomnography-introduction.md
In addition to linking desaturations there may be other quirks with events that either 1) start in sleep and end in wake, or 2) start in wake and end in sleep.
You won't find as many of these discrepancies in newer datasets.