Discrepancy between QC-1.png file and data output from the ActiPASS_Daily_LongFormat_MasterFile.csv

[JanisFie]

I want to screen my data for participants with more than 16h lying down/sleeping. While doing so, I noticed a discrepancy between the QC-1.png and the data output from the ActiPASS_Daily_LongFormat_MasterFile.csv in the parameter "Lie". As you can see in the png and screenshot of the table, the parameter Lie (hrs) is far higher in some cases in the QC-1 file than the parameter Lie (min) in the table.
I have 2 Questions:

• if I want to have the total time a person is lying, do I need to sum up the Variable Lie and the Variable SleepInterval? This would be my best guess based on the ActiPASS_Variable-definitions.
• do you have any explanation for the discrepancy described above? Is Lie differently defined in the QC-1 file than in the table output?

ActiPASS_BatchSettings.txt

[pasan-h]

Hi, I think this is a definition problem 😃. Please carefully read the definition of variables. The quality check figures are preliminary and for quality checks only. They show Lie as a posture within the full 24h. While the variables in the final tables have different definitions and most are defined within the awake time. See the relationship of Duration, NonWear, ValidDuration, SleepInterval, Awake, and all activities in the variable definitions (shown in dark red font).

https://view.officeapps.live.com/op/view.aspx?src=https://github.com/Ergo-Tools/ActiPASS/raw/main/docs/Extra/ActiPASS_Variable-definitions_NewFormat_ActiPASS_Version_1.50.xlsx

[JanisFie]

Thanks for the fast reply!
I assumed as much. However, since I wanted to do the quality check for a lying duration of more than 16 hours/day, I was a bit confused with the quality check figure.

[pasan-h]

At the moment ActiPASS 24h tables follow ProPASS Consortium 24h construct where sleep interval is considered as one big chunk and we don't expose what happens within that. In the future we perhaps can have an option to switch to a more traditional 24h construct where postures are shown instead for full 24h like in the quality check figures.
If you want exact lying time you can still find that if you parse the ActiPASS intermediate 1s epoch data structure via a script.
ActiPASS can export such 1s epoch classification to a CSV.