About physical activity intensity classification

256 views
Skip to first unread message

Jairo Hidalgo Migueles

unread,
Feb 8, 2017, 3:09:11 AM2/8/17
to R package GGIR
Hello,

I have just a short experience analyzing data with GGIR. I have a doubt about day- and night-periods differentiation. I mean, If I want to calculate time spent in each physical activity intensity, it is probable that night-time is classified as sedentary time due to the low movement during this period. Is there any way to avoid this?

Previously I analyzed data on ActiLife (ActiGraph's software) and I introduced a csv indicating what periods correspond to daytime, so the software classified only the time when the participants are awake and there is less probability to classify sleep as sedentary time.

Is there any way to do this with GGIR?

Thanks in advance,
Jairo

Vincent van Hees

unread,
Feb 8, 2017, 8:00:56 AM2/8/17
to Jairo Hidalgo Migueles, R package GGIR
Hi Jairo,
Yes you can do this with g.part 5 (mode = 5 in g.shell.GGIR), but first you need to process it with part 1 till 4.
Best wishes,
Vincent


--
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+unsubscribe@googlegroups.com.
To post to this group, send email to Rpacka...@googlegroups.com.
Visit this group at https://groups.google.com/group/RpackageGGIR.
To view this discussion on the web, visit https://groups.google.com/d/msgid/RpackageGGIR/d802a995-0a3c-4acd-a75f-7c881220f18d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Leo Westbury

unread,
Jul 22, 2020, 7:50:28 AM7/22/20
to R package GGIR
Hello,

I would be very grateful if you could assist with a related query.

If qwindow=c(0,24) is used, are metrics reported in part2_summary.csv (such as ENMO_fullRecordingMeanAD_mean_ENMO_mg_0-24hr and AD_MVPA_E5S_B5M80%_T100_ENMO_0-24hr) averaged over the 24 hour periods, including when the participant was asleep at night?

I'm aware that ENMO_fullRecordingMean uses all data from when the monitor was worn whereas the  _0-24hr variables only used information from valid days (default option of >16 valid hours).


The paper https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5377945/ only used data over the day (periods between waking and sleep onset) for analysis of physical activity. If I wanted to implement a similar approach, are the variables ACC_day_mg_pla,  dur_day_MOD100_400_min_pla and dur_day_VIG400_min_pla from part5_personsummary.csv the correct variables to use for ENMO and minutes in moderate and vigorous activity levels which are averaged over the days and exclude time asleep at night?



I have not used a sleep log so I have used def.noc.sleep=c(1) in the function g.shell.GGIR(). I have set acc.metric = "ENMO".

Many thanks for your assistance.

Best wishes,
Leo


On Wednesday, February 8, 2017 at 1:00:56 PM UTC, Vincent van Hees wrote:
Hi Jairo,
Yes you can do this with g.part 5 (mode = 5 in g.shell.GGIR), but first you need to process it with part 1 till 4.
Best wishes,
Vincent

On 8 February 2017 at 09:09, Jairo Hidalgo Migueles <jairo.hida...@gmail.com> wrote:
Hello,

I have just a short experience analyzing data with GGIR. I have a doubt about day- and night-periods differentiation. I mean, If I want to calculate time spent in each physical activity intensity, it is probable that night-time is classified as sedentary time due to the low movement during this period. Is there any way to avoid this?

Previously I analyzed data on ActiLife (ActiGraph's software) and I introduced a csv indicating what periods correspond to daytime, so the software classified only the time when the participants are awake and there is less probability to classify sleep as sedentary time.

Is there any way to do this with GGIR?

Thanks in advance,
Jairo

--
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 Rpacka...@googlegroups.com.

Vincent van Hees

unread,
Jul 29, 2020, 2:08:16 PM7/29/20
to Leo Westbury, R package GGIR
Hello Leo,

Yes, your understanding is correct. You will need to use the variables from part 5 for that.

Note that you are referring to the old variable names. In GGIR 2.0-0 the part 5 variables names have been much improved. They now better clarify whether they reflect bouted, unbouted, or total time spent in a certain behavioural class.

Vincent

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
To unsubscribe from this group and stop receiving emails from it, send an email to RpackageGGIR...@googlegroups.com.

Leo Westbury

unread,
Aug 1, 2020, 3:45:29 AM8/1/20
to R package GGIR
Hi Vincent,

Many thanks for this information - that's very helpful to know.

Is there a resource that specifies the definitions of the variables generated in the summary.csv files?

Kind regards,
Leo
To unsubscribe from this group and stop receiving emails from it, send an email to Rpacka...@googlegroups.com.

Vincent van Hees

unread,
Aug 3, 2020, 3:50:40 AM8/3/20
to Leo Westbury, R package GGIR
If there is anything missing or unclear then let me know.

Vincent


‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
To unsubscribe from this group and stop receiving emails from it, send an email to RpackageGGIR...@googlegroups.com.

Chan Chan

unread,
Feb 23, 2024, 1:03:56 AM2/23/24
to R package GGIR
Hi,
  I have the same questions here. So, the MVPA was calculated by  dur_day_MOD100_400_min_pla plus  dur_day_VIG400_min_pla. Is that correct?

Best,
chan chan

Leo Westbury 在 2020年7月22日 星期三晚上7:50:28 [UTC+8] 的信中寫道:

Vincent van Hees

unread,
Feb 26, 2024, 10:56:47 AM2/26/24
to Chan Chan, R package GGIR
That depends on which MVPA estimate from the GGIR you are using and how you configured the related input parameters.

If you are using MVPA estimates based on bouts then MOD + VIG is not necessarily the same as MVPA in a bout, unless you do not allow for gaps in bouts and the minimum bout duration equals the epoch length.

Best,
Vincent
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/34126d86-1042-450c-8e68-04066dc7d017n%40googlegroups.com.

Chan Chan

unread,
Feb 28, 2024, 1:30:20 AM2/28/24
to R package GGIR
Thanks, Vincent. Now I understand that. So the average daily total duration of MVPA  is calculated by summing the values of    "dur_day_MVPA_bts_10_min_pla", "dur_day_MVPA_bts_5_10_min_pla" and "dur_day_MVPA_bts_1_5_min_pla" right? 
I put “ boutdur.mvpa = c(1,5,10)” in part five.


Vincent van Hees 在 2024年2月26日 星期一晚上11:56:47 [UTC+8] 的信中寫道:

Chan Chan

unread,
Feb 28, 2024, 2:32:33 AM2/28/24
to R package GGIR

Or, which variable defines minutes in moderate-to-vigorous physical activity per day?
Chan Chan 在 2024年2月28日 星期三下午2:30:20 [UTC+8] 的信中寫道:

Vincent van Hees

unread,
Mar 11, 2024, 4:51:38 AM3/11/24
to Chan Chan, R package GGIR
Hi Chan Chan,

So the average daily total duration of MVPA is calculated by summing the values of  "dur_day_MVPA_bts_10_min_pla", "dur_day_MVPA_bts_5_10_min_pla" and "dur_day_MVPA_bts_1_5_min_pla" right?
I put “ boutdur.mvpa = c(1,5,10)” in part five.

Yes, if you want to only consider bouted behaviour. Another option is to specify boutdur.mvpa = 1 as that will produce just a single estimate of time in MVPA such that you do not have to sum them.

Best wishes,

Vincent

Reply all
Reply to author
Forward
0 new messages