I'm performing an automatic sleep stage classification task using machine learning on SHHS 2.
I can't find if there are some PSG that come from the same patient, because to perform this task there can't be data coming from the same patient both in training and test dataset.
Reading the documentation it seems that each PSG(in SHHS2) come from a different patient, It's is right?
Thank you very much,
You are correct, each PSG in SHHS2 comes from a different patient. There are 2,651 EDFs in this folder (https://sleepdata.org/datasets/shhs/files/polysomnography/edfs/shhs2) and they come from 2,651 different individual research subjects.
Thanks for checking out the site!
Thank you very much for your reply!
Sorry for the inconvenience,
I'm looking for whether SHHS has one variable that identifies the quality of sleep, so that my model, in addition to estimating the hypnogram, can also retain this value.
Do you know if it is present? Within the "variables" section I can't find it.
Thanks a lot,
Were you thinking of self-reported sleep quality? Maybe these variables from the Morning Survey in SHHS2 would help: https://sleepdata.org/datasets/shhs/variables?folder=Questionnaires%2FSHHS2%2FMorning+Survey
If you had another sort of quality in mind let me know and I can comment again.
Thank you for the reply,
What I was looking for is a "medical" variable that assesses the quality of sleep.
If it were not present I could opt for the linked one related to the morning survey, but not being "medical" I would prefer to avoid it.
Hmm, I still don't fully understand. These were research subjects, so there won't be any sort of subjective "interpretation" (by a medical professional).
The polysomnologist who scored the study gave ratings of the quality of individual signals and the overall quality of the sleep study, e.g. https://sleepdata.org/datasets/shhs/variables?folder=Measurements%2FPolysomnography%2FSignal+Quality
Sleep efficiency (https://sleepdata.org/datasets/shhs/variables/slp_eff) is a commonly used metric to assess the "quality" of a person's sleep. This is the proportion of time spent asleep within the time in bed. Higher efficiency is thought to often indicate a "better" night's sleep.
Thanks so much for the explanation.
Sleep efficiency could be the metric I was looking for.
Thanks again for the help.
I apologize but I have to disturb one last time.
Always regarding SHHS2, I am looking for a variable that indicates the patient's health status.
I found ql201 (Quality of Life: General health), but it is taken (if I understand correctly) from the answers to the questionnaire given by the patients.
So is there an equal variable but assigned by a doctor?
I give an example: "state of health of the patient with id X: healthy / unhealthy"
No, there isn't a variable like that. These were all research studies, so neither the patient nor their sleep study were judged/assessed by a health professional.