Extracting non-wear in minute-by-minute epochs

[Anastasiya S]

Dear Vincent and team, 

First of all, thank you for putting together such an incredible and comprehensive package. It has been a pleasure to use so far, and to read the associated papers.

I was wondering whether it is possible to extract non-wear scoring in minute-by-minute epochs? (or anything that could ultimately be aggregated into 60s epochs. 

I am hoping to look at some additional analyses that your package doesn't address, and I don't think it makes sense to use 2 different non-wear algorithms for the same set of analyses (e.g. yours and Troiano or Choi's). 

Thank you kindly for your help,

Anastasiya 

[Vincent van Hees]

Hi Anastasiya,

With argument windowsizes you can change the resolution of the non-wear detection from 15 minutes to 1 minute. See also package documentation.

The non-wear scores are stored in .RData files in output subfolder meta/ms2.out. Inside these you will find object IMP$rout this is a matrix with all the labels where you probably only want to work with column 5, see also documentation for 'rout'. Corresponding timestamps are in IMP$metalong.

Do you think the additional analyses you want to do should be added to GGIR? I am happy to be available for a chat to discuss options.

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/4f4ced11-1339-4a77-90ef-eb9e02700597n%40googlegroups.com.

[Anastasiya S]

Dear Vincent,

Thank you kindly for your prompt and informative response! Which method of imputation is used by GGIR?

I’m interested in looking at light exposure within actigraphy data, which I don’t believe can be done with the GGIR package currently.

Thank you again for all of your help,

Anastasiya

On Oct 8, 2020, at 14:44, Vincent van Hees <v.va...@accelting.com> wrote:

Hi Anastasiya,

With argument windowsizes you can change the resolution of the non-wear detection from 15 minutes to 1 minute. See also package documentation.

The non-wear scores are stored in .RData files in output subfolder meta/ms2.out. Inside these you will find object IMP$rout this is a matrix with all the labels where you probably only want to work with column 5, see also documentation for 'rout'. Corresponding timestamps are in IMP$metalong.

Do you think the additional analyses you want to do should be added to GGIR? I am happy to be available for a chat to discuss options.

Vincent

Dr. Vincent van Hees

Independent consultant

https://accelting.com/

<ACCELTING.png>

[Vincent van Hees]

Hi Anastasiya,

Light is stored by GGIR in the same IMP$metalong object I referred to below. It would be relatively easy to embed some summary statistics on light in the day or person-level summary. You can either develop this yourself and make a pull request (if you are a programmer) or I am happy to do this as a consultancy project.

For imputation see package documentation and most GGIR papers. We impute per timepoint by taking the average across similar time points on other days of the recording.

Best,

Vincent

Dr. Vincent van Hees

Independent consultant

https://accelting.com/

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

[Alex Li]

Hi Vincent and  Anastasiya,

Thanks for sharing your experience about the extracting nonwearing time epoch.

I am focusing on the same task. 2 questions from my group:

Q1: Could we take the epoch with r1 = 1 as non-wear time? From your document I noticed ' Columns 1 = monitor non wear' . And

Q2: What are the differences of r1 and r3 ( column 3 = additional nonwear )? In the attached case, the epoch of 00:15-00:30 was not labelled as non-wearing.  

Many thanks.

Regards,

Alex

[Vincent van Hees]

Hi Alex,

GGIR does the non-wear detection in two stages as described in papers and documentation. The two columns reflect the two stages.

Would you mind clarifying why you need this information? If you would like this to be part of the timeseries csv output produced by GGIR part 2 then please indicate this as that would allow all GGIR users to benefit from what you are trying to do.

Vincent

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

To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/9e23c11f-6c1a-4747-b242-421f8951cacdn%40googlegroups.com.