. 2024 Sep 12;10:20552076241277150. doi: 10.1177/20552076241277150

Table 1.

Sleep variables included in the main analysis comparing data from the Garmin Vivofit 4 against polysomnography (PSG).

Variable	Description	Equivalence Zone^a
Sleep onset (bedtime)	Time of falling asleep	± 0.5 hours
Sleep offset (risetime)^b	Time of waking up	± 0.5 hours
Sleep onset latency^c	Amount of time in bed before falling asleep, in minutes	± 15 minutes
Total sleep time (TST)^d	Amount of time spent sleeping, in hours	± 0.5 hours
Wake after sleep onset (WASO)	Amount of time spent awake after falling asleep, in minutes	± 15 minutes
Sleep efficiency^e	TST as a percentage of total time in bed	± 5%
Light sleep^f,g	Percent of sleep spent in light stages	± 5%
Deep sleep^f,h	Percent of sleep spent in deep stages	± 5%

^{^a}

For equivalence testing, this was used as the acceptable margin of error, within which a group-level estimate (i.e. mean) from the Garmin could be considered equivalent to PSG.

^{^b}

For PSG, given as “lights on” time, also equal to the sum of bedtime, TST, and WASO; for Garmin, calculated using the latter sum.

^{^c}

Garmin value calculated as the time from “lights out” (PSG) to bedtime, with negative values excluded from further analysis (n = 1).

^{^d}

Garmin value calculated as the sum of light and deep sleep time.

^{^e}

Garmin value calculated by dividing TST by total in-bed time (i.e. latency + TST + WASO).

^{^f}

Raw PSG values were expressed in percentage units; Garmin percentages were calculated by dividing minutes in each stage by TST.

^{^g}

For PSG, defined as either stages N1–N2 plus rapid eye movement (REM Light) or only stages N1–N2 (REM Deep).

^{^h}

For PSG, defined as either stage N3 (REM Light) or stage N3 and rapid eye movement (REM Deep).