How to derive Non-wear and Nap time windows from output

[Adithya Palle]

Hi all,

I am trying to find sleep windows, nap windows, and non-wear time windows using GGIR. So far, I have only been able to find the sleep windows by looking at the sleep onset columns in the part 5 output. 

Is there a similar time recording for non-wear and napping times, or is that data only available as proportions of the total day?

Thank You,

Adithya 

[Vincent van Hees]

Hi Adithy,

GGIR does not provide nap estimates, but reports time spent in sustained inactivity bouts (abbreviated as sib). I intentionally named it like that because there is no evidence that these reflect naps. For example, they could also reflect brief episodes of sensor non-wear.

Unfortunately, authors like https://pubmed.ncbi.nlm.nih.gov/34637446/ and https://www.nature.com/articles/s41467-020-20585-3 ignored that and referred to sibs as 'naps', because that makes it easier to score publications or something, but it is not evidence based

We do have some prototype code for nap detection in 3-4 year olds based on waist worn accelerometer data, which I would not recommend you to use and which I therefore have not documented well yet. 

Best, Vincent

Dr. Vincent van Hees | Independent consultant | https://accelting.com/

------- 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/cf78a4c0-0d96-48d4-b13a-ec43ffbeb6b8n%40googlegroups.com.

[Adithya Palle]

Hi Vincent,

Thanks for the response, I will take your advice and not attempt to record "nap" windows.

Are there any concrete time recordings for non-wear windows available from the GGIR output? I assumed there was because the part 5 output produces data such as "nonwear_perc_day" which measures the percentage of non-wear time during the day. If there are, where would I have to look to find them?

Thank You,

Adithya 

[Vincent van Hees]

Are there any concrete time recordings for non-wear windows available from the GGIR output? I assumed there was because the part 5 output produces data such as "nonwear_perc_day" which measures the percentage of non-wear time during the day. If there are, where would I have to look to find them?

​

See documentation on time series output files:

https://cran.r-project.org/web/packages/GGIR/vignettes/GGIR.html#571_Time_series_output_files

Cheers,

Vincent

Dr. Vincent van Hees | Independent consultant | https://accelting.com/

------- Original Message -------

To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/39d06ca8-89a9-4687-afd0-bf91e862d29dn%40googlegroups.com.

[Adithya Palle]

Hi Vincent,

Thanks for the help.

I was able to find the time series output file, but I am having another issue where the first and last days of my data are not being included in this time series or any of the part 5 outputs (because they are partial days). I read in the documentation that this is an intended effect, but that I could remove this feature by adjusting the "minimum_MM_length.part5" parameter. 

I've tried this, but have had no success with getting the output for these first and last days.

```

g.shell.GGIR(#=======================================

# INPUT NEEDED:

mode=c(1,2,3,4,5),

datadir= c(locs[1]),

outputdir= locs[2],

studyname=id,

f0=1,

overwrite = FALSE,

# sample_rate = 80,

#-------------------------------

# Part 1:

#-------------------------------

# Key functions: reading file, auto-calibration, and extracting features

chunksize=1, printsummary=TRUE,

#-------------------------------

# Part 2:

#-------------------------------

strategy = 1,

hrs.del.start = 0, hrs.del.end = 0,

# winhr = c(5,10),

# ilevels = c(seq(0,400,by=50),8000),

mvpathreshold =c(100,120),

bout.metric = 4,

closedbout=FALSE, dofirstpage =FALSE,

#-------------------------------

# Part 3:

#-------------------------------

# Key functions: Sleep detection

ignorenonwear = FALSE,

#-------------------------------

# Part 4:

#-------------------------------

do.visual = TRUE,

#-------------------------------

# Part 5:

# Key functions: Merging physical activity with sleep analyses

#-------------------------------

threshold.lig = c(30), threshold.mod = c(100), threshold.vig = c(400),

boutcriter = 0.8, boutcriter.in = 0.9, boutcriter.lig = 0.8,

boutcriter.mvpa = 0.8, boutdur.in = c(1,10,30), boutdur.lig = c(1,10),

boutdur.mvpa = c(1), save_ms5rawlevels = TRUE,

do.sibreport = TRUE, frag.metrics="all", ,part5_agg2_60seconds=FALSE,

timewindow = c("MM", "WW"), minimum_MM_length.part5 = 1,

#-----------------------------------©

# Report generation

#-------------------------------

# Key functions: Generating reports based on meta-data

do.report=c(2,4,5),

visualreport=TRUE, # dofirstpage = TRUE,

viewingwindow = 1 )

}

```

Do you have any ideas on how to fix this so I can get time series/summary data for the first and last day as well?

Thank you so much,

Adithya Palle

[Vincent van Hees]

Does argument 'save_ms5raw_without_invalid' offer a solution? Possibly you then also do not need to use `save_ms5raw_without_invalid`. I have not tried this myself, so would be good to hear back from you how it goes.

Thanks,

Vincent

------- Original Message -------

To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/9a74676c-c916-4325-8480-2a8ccd7ac86cn%40googlegroups.com.

[Adithya Palle]

Hi Vincent,

This worked! The time series output now does include the clipped-off days at the start and end. The day-summaries in part 5 still exclude these days, but that is fine as I should be able to find what I want in the time series output.

Thank You,

Adithya Palle

[Adithya Palle]

Hi Vincent,

Thanks again for the help with the invalid day issue.

I am also working with some data with less than 2 nights of sleep, and I was hoping the 'save_ms5raw_without_invalid' argument would help me read that, but I keep getting an error:

Error in `[.data.frame`(daysummary_tmp, , MainMetric) : 

  object 'MainMetric' not found

Calls: g.shell.GGIR ... <Anonymous> -> g.plot5 -> which -> [ -> [.data.frame

Execution halted

It seems to be the same issue that is discussed in this thread :

https://github.com/wadpac/GGIR/issues/192

Do you know of any arguments in GGIR that can be tweaked to allow me to read time series data for data with only 1 or 2 days?

Thank you for your help,

Adithya

[Vincent van Hees]

Are you using a different acceleration metric then the default? If yes, then specify argument acc.metric to tell GGIR that that is the accelerometer to be used for the visualreport.

Best, Vincent

Dr. Vincent van Hees | Independent consultant | https://accelting.com/

------- Original Message -------

To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/2d29f64e-83b2-431f-b713-e1d9b2d17814n%40googlegroups.com.