GGIR seems to be unable to classify multiple sleep periods on one day.

37 views
Skip to first unread message

Romy Stuik

unread,
Sep 24, 2021, 7:02:49 AM9/24/21
to R package GGIR
Hi all, 

I am graduating on studying the effects of nightshifts on mental health.
Now that i am studying the results, the results summary shows clearly that GGIR seems to be unable to correctly classify when people have multiple sleep periods on one day.

I am attachting one file which clearly shows how multiple sleep periods are not correctly seen by GGIR, leading to unclear data, and incorrect sleep onsets/offsets. 

Is there any way to make GGIR sensitive for multiple sleep periods? Is there another outcome data that does reliably show me the exact sleep times? Or could I manually classify the correct sleep periods?

In other words: how could i handle this kind of outcome to correctly get the sleep data?

many thanks in advance,
greetings,
Romy Stuik
Report_69702_0000000022.cwa.pdf

Vincent van Hees

unread,
Sep 24, 2021, 10:31:16 AM9/24/21
to Romy Stuik, R package GGIR
Hi Romy,

You may want to have a look at the .csv-reports first, because those are core results, which have been tested extensively in datasets with irregular sleep patterns.

The pdf visualisation is an attempt to visualise the quantitative results, but especially when sleep is irregular the timing of sleep window isn't always shown correctly, which is something that needs improving. For the time being it is probably good to treat the timing of sleep in the pdf visualisation with caution.

GGIR tries to detect the largest sleep period in the day. I know this doesn't fit well with the scenario where people have irregular sleep and end up with multiple main sleep periods in the day. However, note that the day time 'sustained inactivity bouts (=sib)'  are reported in the GGIR part 4 csv-report. I am trying to be careful with calling them naps as whether they are naps needs more research. You can use these to get a sense of overall time spend in behaviours that look like sleep/resting by adding up sleep in the sleep period time window (nocturnal sleep) and daytime resting (=those sib periods).

Vincent

Dr. Vincent van Hees
Independent consultant

ACCELTING.png


‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
--
You received this message because you are subscribed to the Google Groups "R package GGIR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to RpackageGGIR...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/c9a2418e-3350-4cf0-a169-2efafbe1606fn%40googlegroups.com.

Ori - Roee Admon Lab

unread,
Oct 23, 2021, 10:13:42 AM10/23/21
to R package GGIR
Hello everyone,
I have noticed that I cannot get those 1st page reports Romy got on the file attached.
Of course my shell configuration includes dofirstpage=TRUE, visualreport=TRUE.

Can someone share the exact argument for these visual graphs?
Thanks, Ori

Vincent van Hees

unread,
Oct 24, 2021, 3:46:33 PM10/24/21
to Ori - Roee Admon Lab, R package GGIR
First page requires more than 1 day and 1 night of valid data. Maybe your data does not meet that criteria?

Next time, please share your g.shell.GGIR function call and any other information you think could be relevant for others to understand your issue.

Vincent

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
Reply all
Reply to author
Forward
0 new messages