Support Center

What does the "Detect Sleep Periods" button do and how does it work?

Last Updated: Mar 02, 2018 10:08AM CST

What is the "Detect Sleep Periods" button in Sleep Analysis?

Note: Prior to ActiLife 6.12 this button was labeled: "Auto Score"

The "Detect Sleep Periods" feature in the ActiLife sleep tool automatically estimates Time in Bed (TIB) and Time Out of Bed (TOB) for a given dataset by analyzing the data and looking at activity levels and sleep/wake calculation. To use the tool, load a dataset in the Sleep Analysis tool and click the "Detect Sleep Periods" button.  

2015-07-28_10-22-37.png

 

IMPORTANT: The "Detect Sleep Periods" tool will only work on datasets collected using the ActiSleep, ActiSleep+ or wActiSleep+ devices.  If you wish to use the Auto Score feature on a dataset that has not been captured using an ActiSleep device, the feature can be enabled remotely.  Contact ActiGraph at sales@theactigraph.com for pricing.

 

How does the "Tudor-Locke" auto sleep period detection algorithm work?

Using the Tudor-Locke paper Fully automated waist-worn accelerometer algorithm for detecting children’s sleep period time separate from 24-hour physical activity or sedentary behaviors, we have derived an algorithm to calculate bedtimes automatically. Tudor-Locke uses the results calculated by either Cole-Kripke or Sadeh algorithms (which classify epochs as sleep vs. wake), and then applies its own formulas to determine other variables. 

Parameter

Tudor-Locke Paper

Tudor-Locke Implementation in ActiLife

Calculating each epoch as sleep or awake

Combines Sadeh sleep algorithm and inclinometer results to classify each epoch. An epoch is only asleep if BOTH the Sadeh algorithm says the subject is asleep and the inclinometer value is OFF.

Uses currently selected sleep algorithm to classify each epoch as asleep or awake. Ignores inclinometer values.

Wear position of device

Algorithm is based off of waist-worn device.

Not specified.

Age of subject

Developed on fourth-grade schoolchildren.

Not specified.

Bedtime definition

5 consecutive asleep minutes.

Default value is 5 minutes.
It can be customized.

Wake time definition

First 10 consecutive minutes of awake time following a period of sleep.

Default value is 10 minutes.
It can be customized.

Minimum amount of time between bedtime and wake time

Must be at least 160 minutes between in bed and out of bed time.

Default value is 160 minutes.
It can be customized.

Multiple sleep periods in a day

Allowed

Allowed

Maximum sleep period length

None

1440 minutes. ActiLife does not allow sleep periods longer than 24 hours.

Auto_Sleep_Period_Detection_Options_2015-07-28_10-20-02.png

 

A few notes about the Tudor-Locke algorithm in ActiLife:
  • ActiLife requires a 60-second epoch file to run the Tudor-Locke auto sleep period detection
  • Due to the algorithm requiring several minutes of  'sleep' epochs at the start of a sleep period, the latency will be zero (0) for any sleep periods that are found.
     

How does the "ActiGraph" auto sleep period detection algorithm work?

The Auto Score feature works by analyzing the activity levels using minute-by-minute data, similar to how Wear Time Validation identifies periods of non-wear.  Below is a description of how the Auto Score feature works.

 

Using Troiano Wear Time Validation Algorithm, the ActiGraph auto sleep period detection method looks for non-wear bouts with the following parameters:     
     
 A. 60s Epoch Length
  • Must be at least 30 minutes long.
  • No threshold.
  • Spike Tolerance of 90 minutes.
  • First and Last 10 epochs of data must be zeros.
  • If any wear period is less than 10 minutes, add it to adjoining non-wear period.
B. 30s Epoch Length
  • Must be at least 30 minutes long.
  • No threshold.
  • Spike Tolerance of 70 minutes.
  • First and Last 10 epochs of data must be zeros.
  • If any wear period is less than 10 minutes, add it to adjoining non-wear period.

C. 10s Epoch Length
  • Must be at least 30 minutes long.
  • No threshold.
  • Spike Tolerance of 50 minutes.
  • First and Last 30 epochs of data must be zeros.
  • If any wear period is less than 2 minutes, add it to adjoining non-wear period.


Once we have found the non-wear periods, we ignore:
  • Non-wear periods greater than 24 hours - this eliminates long periods of time that aren’t sleep.
  • Non-wear periods that have almost all zeros (must have more than 5 epochs or more of non-zeros) - this eliminates time where the device is most likely not being worn.


The non-wear periods that are left are turned into the bed times.

 

Contact Us

84597daf644fbf443ca13c8a01998d15@actigraph.desk-mail.com
https://cdn.desk.com/
false
desk
Loading
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
about
false
Invalid characters found
/customer/en/portal/articles/autocomplete