mrueschman

Thanks, you're right, I should have clarified further. The "immobile time" counting is done on a continuous basis by the Actiware software, so if there were 5 minutes of immobile time before the beginning of a REST interval this can result in a value of 0 minutes of sleep latency.

I think the point I was trying to get at is that sleep latency in actigraphy was most commonly in the 0-10 minute range because the scoring rules called for REST intervals to be set approximately around the time activity counts started to decrease (in other words, around the time the subject started becoming more immobile than mobile). People are typically lying still when they start to try and fall asleep, yet (of course) the actigraphy device can't tell when the person truly transitions from wake to sleep.

The scorers did reference self-reported sleep diaries in the setting of REST intervals, however this was just one input used (others being: activity count levels; light levels; event markers). Perhaps if the sleep diary alone was used in the setting of REST intervals we might get "truer" estimates of sleep onset latency. For instance, a subject reports they got in bed and tried to sleep at 10 p.m., yet the actigraphy device still shows sporadic movement until 10:30 p.m., when the subject triggered the "5 minutes of immobile time" for sleep onset, thus giving ~30 minutes of sleep onset latency for that night.

Bottom line: sleep latency and actigraphy is very tricky!

Hey - thanks for checking in again.

I think we communicated before about your 2nd question (https://sleepdata.org/forum/mnc/1#comment-1066). I would check out the .STA files here: https://stanfordmedicine.app.box.com/s/r9e92ygq0erf7hn5re6j51aaggf50jly/folder/53209541138

I will ping the MNC data contributor about your first question regarding the SSC subjects who do not appear in the cohorts_deid file.

Good question - sleep onset latency is difficult to discern from actigraphy alone. The reason it is ~5 minutes in MESA actigraphy data is because the Actiware scoring software used a sleep onset detection method of "5 minutes of immobile time". Scorers generally set bedtime (i.e., the start of an Actiware REST interval) around the time the activity signal diminished, which typically corresponded to what ActiWare considers "immobile time" (activity counts < 2 at 30-second epochs).

We don't have any additional information to provide beyond what is available on NSRR.

Another approach to consider is to recompute onset latency from the epoch-by-epoch data using different criteria.

Hey Ethan - I sent this via email, but posting here as well!

Thanks for your interest in the resource. The SHHS baseline (Visit 1 / "index") visits took place around ~1995. Some of the cohorts tracked incident CVD outcomes up until ~2010, so most subjects have somewhere between 10-15 years of follow-up tracking those outcomes (e.g., https://sleepdata.org/datasets/shhs/variables/censdate).

The second home PSG at SHHS Visit 2 appears to have occurred ~5-7 years after the index PSG (e.g., https://sleepdata.org/datasets/shhs/variables/stdatep).

The SHHS data contributor provided some documentation of the analytic database, which appears to contain some information on any_cvd coding: https://sleepdata.org/datasets/shhs/files/m/browser/documentation/SHHS_Documentation_of_Analytic_Database.pdf?inline=1

It looks like another Forum user also asked a similar question before: https://sleepdata.org/forum/cvd-variable/

The .annot files are now available here: https://sleepdata.org/datasets/nchsdb/files/sleep_data

Thanks for using the site!

1 technician scored each record. If you wish to stratify by scorer, you could use this variables: https://sleepdata.org/datasets/mesa/variables/scorerid5

The team of PSG scorers routinely completed reliability exercises, which are described briefly in this documentation: https://sleepdata.org/datasets/mesa/files/documentation/MESA_Sleep_Data_Documentation_Guide.pdf

Thanks for using the site!

Hey Yan --

Unfortunately, there are known issues with the desaturation annotations in SHHS. From this page:

Oxygen desaturation events from the annotation files may be absent and/or may not line up as anticipated with respiratory events. Original scoring data were indelibly changed when converting to a newer version of the scoring software that allowed for the creation of EDF/XML files. Users may opt to derive new desaturation events from the SaO2 signal.

By the way, I know you have posted on the forum in the past about EDF/annotation software you were developing. Is that where your screenshot is from? It looks great! Is the software available for the general public to download/use? Thanks!

Thanks for checking out the site. Here are some insights about ahi_a0h3 that may help:

1. The denominator is the sum of all epochs of scored sleep.
2. Watch out for respiratory events scored in wake epochs, as these would not have been counted in the numerator. Note that the scoring procedures were meant to remove all events scored in wake, however a handful may have remained in the final scoring output.
3. Both the Hypopnea and Unsure event concept tags represent hypopneas (see note here: https://sleepdata.org/datasets/mesa/pages/polysomnography-introduction.md), however you need to ensure you are only counting hypopneas associated with a >=3% SpO2 desaturation. Hypopneas were scored regardless of the associated desaturation (or lack thereof), so if you are counting all hypopneas this will likely give inflated AHI estimates. There are SpO2 desaturation event tags in the files too, however I do not know the specifics about how the Compumedics Profusion software linked these to respiratory events (e.g., time windows for events to occur to be considered linked).

Hope this helps - let me know if you have follow-up questions!

Hello, thanks for your question!

1. All of these cohorts were scored by trained technicians. Raw signals and scoring annotations are available for users to download.
2. Yes, the signal data are provided as .EDF (https://www.edfplus.info/). The scoring annotations come in different formats depending on the dataset.
3. It looks like the SHHS and MESA HRV projects include many of the metrics you listed. More information is available on these page: https://sleepdata.org/datasets/shhs/pages/13-hrv-analysis.md and https://sleepdata.org/datasets/mesa/pages/hrv-analysis.md

The dataset documentation contains information on the scoring procedures and available PSG files, e.g.:

• SHHS

  • https://sleepdata.org/datasets/shhs/pages/mop/6-00-mop-toc.md
  • https://sleepdata.org/datasets/shhs/pages/05-polysomnography-introduction.md

• MESA

  • https://sleepdata.org/datasets/mesa/files/documentation/MESA_Sleep_Polysomnography_Scoring_Manual.pdf
  • https://sleepdata.org/datasets/mesa/pages/polysomnography-introduction.md

• WSC

  • https://sleepdata.org/datasets/wsc/files/documentation/WSCS_Manual_of_Operations.pdf
  • https://sleepdata.org/datasets/wsc/pages/polysomnography-introduction.md

I did some googling and my best guess would be ventricular ectopic (VE) and supraventricular ectopic (SVE). Thanks for checking out the site!