Skip to main content

Table 5 Criterion and convergent validity of accelerometer-based methods for preschoolers, sorted by methodological study quality, quality of evidence (level of evidence), and study result rating

From: Systematic review of accelerometer-based methods for 24-h physical behavior assessment in young children (0–5 years old)

  Study Study populationa Outcome(s) & Setting Comparison measure(s) Method
ological study qualityb
Placement Device-based methodc Results Study result ratingd Quality of evidencee
(level of evidence)
Criterion validity Adolph et al. (2012) [51] n = 64
Age = 4.5 ± 0.8 (3 to 5) years
Sex = 42.2% girls
cpm
Laboratory (structured activities)
Indirect calorimetry using breath-by-breath VO2 and VCO2 method to determine AEE VG Chest Ah-cts-uni-60 r = .72*** +  Moderate
Right hip RT-cts-VM-60 r = .80** + 
RT-cts-VA-60 r = .80**  + 
RT-cts-HA-60 r = .78** + 
RT-cts-DA-60 r = .74** + 
Zakeri et al. (2013) [55] n = 69
Age = 4.6 ± 1.0 (3 to 5) years
Sex = 50.7% girls
Minute-by-minute EE
Laboratory (structured activities)
Indirect calorimetry using breath-by-breath VO2 and VCO2 to determine EE VG Right hip AG-cts-VA-60 r = .76*** +  Moderate
AG-cts-HA-60 r = .74*** + 
AG-cts-DA-60 r = .76*** + 
AG-CSTS-VA/HA/DA/VM-60 MAPE = 0.3 ± 6.9%, CCC = .93 + 
AG-MARS-VA/HA/DA/VM-60 MAPE = 0.3 ± 4.8%, CCC = .95 + 
Bélanger et al. (2013) [66] n = 12
Age = 3.1 ± 1.0 (2 to 5) years
Sex = 66.7% girls
Sleep quality: sleep latency, TST, WASO (number), and sleep efficiency
Free-living
(at home)
Polysomnography to score sleep stages as sleep and wake VG Non-dominant wrist
(raw data)
Aw-AS-VA-30 Sleep latency: ICC = .96**
TST: ICC = .94**, overestimated by 13 min**
WASO: ICC = .28, underestimated by 20**
Sleep efficiency: ICC = .76**, overestimated by 2 min*
Se = 98.7%, Sp = 58.7%, accuracy = 95.6%, NVP = 81.0%
+  Low
Aw-ACT40-VA-30 Sleep latency: ICC = .92**
TST: ICC = .94**, underestimated by 25 min**
WASO: ICC = .42, overestimated by 37**
Sleep efficiency: ICC = .73**, underestimated by 4 min**
Se = 92.7%, Sp = 69.9%, accuracy = 90.7%, NVP = 47.4%
± 
Aw-ACT80-VA-30 Sleep latency: ICC = .92**
TST: ICC = .95*
WASO: ICC = .32, overestimated by 25**
Sleep efficiency: ICC = .70**
Se = 95.8%, Sp = 56.7%, accuracy = 92.4%, NVP = 55.0%
± 
Non-dominant wrist
(adjusted data)
Aw-ACT40-VA-30 Sleep latency: ICC = .96**
TST: ICC = .94**, underestimated by 58 min**
WASO: ICC = .28, overestimated by 35**
Sleep efficiency: ICC = .76**, underestimated by 9.4 min**
Se = 87.9%, Sp = 81.0%, accuracy = 87.5%, NVP = 39.4%
+ 
Aw-ACT80-VA-30 Sleep latency: ICC = .96**
TST: ICC = .97**, underestimated by 21.5 min**
WASO: ICC = .36, overestimated by 34**
Sleep efficiency: ICC = .82**, underestimated by 3.5 min**
Se = 93.4%, Sp = 70.9%, accuracy = 91.4%, NVP = 49.6%
+ 
Aw-AS-VA-30 Sleep latency: ICC = .93**
TST: ICC = .98**
WASO: ICC = .36, underestimated by 20**
Sleep efficiency: ICC = .90**
Se = 97.7%, Sp = 61.2%, accuracy = 95.0%, NVP = 72.1%
± 
Left ankle Aw-ACT40-VA-30 Sleep latency: ICC = .83**
TST: ICC = .94**, underestimated by > 25 min**
WASO: ICC = .01, overestimated by 36**
Sleep efficiency: ICC = .80**, underestimated by 6.5 min**
Se = 90.5%, Sp = 75.1%, accuracy = 89.3%, NVP = 41.7%
± 
Aw-ACT80-VA-30 Sleep latency: ICC = .85**
TST: ICC = .95**
WASO: ICC = .09, overestimated by 28**
Sleep efficiency: ICC = .81**
Se = 95.0%, Sp = 65.0%, accuracy = 92.1%, NVP = 53.7%
 ± 
Aw-AS-VA-30 Sleep latency: ICC = .81**
TST: ICC = .91**
WASO: ICC = .14, underestimated by 20**
Sleep efficiency: ICC = .77**
Se = 97.6%, Sp = 57.7%, accuracy = 94.6%, NVP = 76.8%
± 
Roscoe, James, & Duncan (2017) [67] n = 21
Age = 4.7 ± 0.5 (4 to 5) years
Sex = 38.1% girls
Activity intensity: SB (< 2 METs), LPA (2–2.99 METs), and MPA (3–5.99 METs)
Laboratory (structured activities)
Indirect calorimetry using breath-by-breath VO2 and VCO2 to determine EE VG Non-dominant wrist GA-Ro-VM-60 SB: AUC-ROC = .99, 95% CI (.98 to 1.0), Se = 90%, Sp = 90%
LPA: AUC-ROC = .75, 95% CI (.65 to .85), Se = 40%, Sp = 20%
MPA: AUC-ROC = .92, 95% CI (.86 to .98), Se = 86%, Sp = 40%
± 
(SB + ;
PA -)
Low
Dominant wrist GA-Ro-VM-60 SB: AUC-ROC = .99, 95% CI (.97 to 1.0), Se = 100%, Sp = 10%
LPA: AUC-ROC = .76, 95% CI (.60 to .86), Se = 10%, Sp = 85%
MPA: AUC-ROC = .90, 95% CI (.83 to .96), Se = 76%, Sp = 40%
-
(SB ± ;
PA -)
Pfeiffer et al. (2006) [53] n = 18
Age = 4.4 ± 0.7 (3 to 5) years
Sex = 61.0% girls
cpm
Laboratory (structured activities)
Indirect calorimetry using breath-by-breath VO2 VG Right hip Ac-cts-omni-60 Validation
r = .89, rsp = .90, R2 = .96, SE = 3.02, AIC = 435.9, ICC = .59, rsp = .80***
+  Very low
Activity intensity: MPA and VPA
Free-living (in- and outdoor activities at preschool)
D Ac-Pf-omni-60 Cross-validation
MPA: accuracy = 73%, κ = .40, modified κ = .46
VPA: accuracy = 85%, κ = .26, modified κ = .71
PA: ± 
Pate et al. (2006) [52] n = 29
Age = 4.4 ± 0.8 (3.3 to 5.9) years
Sex = 46.2% girls
cpm
Laboratory (structured activities)
Indirect calorimetry using breath-by-breath VO2 and VCO2 to determine EE VG Right hip AG-cts-VA-60 Validation
rp = .82, R2 = .90, SE = 4.70, AIC = 735.8, ICC = .57, rsp = .66***
±  Very low
Activity intensity: MVPA and VPA
Free-living (in- and outdoor activities at preschool)
D AG-Pa-VA-60 Cross-validation
MVPA: accuracy = 69%, Se = 96.6%, Sp = 86.2% κ = .36, modified κ = .38
VPA: accuracy = 81%, Se = 65.5%, Sp = 95.4%, κ = .13, modified κ = .62
PA: -
Sijtsma et al. (2013) [54] n = 30
Age = 3.4 ± 0.3 (3.1 to 4.4) years
Sex = 60.0% girls
Mean activity cpm (ACM) and mean total activity counts per day (ACD)
Free-living (at home for 5 days)
Indirect calorimetry using DLW to determine TEE;
breath-by-breath VO2 and VCO2 to determine SMR to calculate PAL and AEE
D Middle lower back TD-cts-tri-60 PAL: ACM rp = .61**, ACD rp = 46*
- ACM, gender & weight, R2 = .50*; ACD, height, gender & sleep duration, R2 = .48*
AEE: ACM rp = .56*, ACD rp = .38
- ACM, R2 = .31*; ACD, gender & sleep duration, R2 = .39*
TEE: ACM rp = .34, ACD rp = .21
- Multivariate models with ACM & SMR or ACM, ACD & sleep duration not significant
Using 5 or 3 days of data did not provide different results for PAL estimation, MD = -0.02 ± 0.07, LoA (-0.16 to 0.11)
- Very low
Ahmadi et al. (2020) [84] n = 31
Age = 4.1 ± 1.0 (3 to 5) years
Sex = 20% girls
EE
Free-living (free play)
Indirect calorimetry using breath-by-breath VO2 and VCO2 to determine EE A Non-dominant wrist AG-RF-VM-10 Free-living (FL) model: RMSE = 0.63 (SD = 0.47) kcal/min, RMSE = 0.96 (SD = 0.67) METs, MAPE = 27.4% (SD = 14.0)
Retrained lab (RL) model: RMSE = 0.66 (SD = 0.47) kcal/min, RMSE = 1.01 (SD = 0.67) METs, MAPE = 28.3% (SD = 15.0)
Existing lab, FL and RL models: mean bias not different from zero, predicted EE within ± 6% of measured EE
? Very low
AG-SVM-VM-10 Free-living (FL) model: RMSE = 0.64 (SD = 0.51) kcal/min, RMSE = 0.99 (SD = 0.73) METs, MAPE = 25.4% (SD = 12.2)
Retrained lab (RL) model: RMSE = 0.65 (SD = 0.54) kcal/min, RMSE = 0.99 (SD = 0.77) METs, MAPE = 26.0% (SD = 12.6)
Existing lab, FL and RL models: mean bias not different from zero, predicted EE within ± 6% of measured EE
?
Right hip AG-RF-VM-10 Free-living (FL) model: RMSE = 0.63 (SD = 0.42) kcal/min, RMSE = 0.96 (SD = 0.59) METs, MAPE = 28.1% (SD = 12.0)
Retrained lab (RL) model: RMSE = 0.67 (SD = 0.41) kcal/min, RMSE = 1.02 (SD = 0.57) METs, MAPE = 28.3% (SD = 12.7)
FL and RL models: EE overestimated during play sessions with low total EE and underestimated during sessions with high total EE
Existing lab models: mean bias not different from zero, predicted EE within ± 6% of measured EE
?
AG-ANN-VM-10 Free-living (FL) model: RMSE = 0.63 (SD = 0.43) kcal/min, RMSE = 0.96 (SD = 0.61) METs, MAPE = 27.1% (SD = 11.1)
Retrained lab (RL) model: RMSE = 0.65 (SD = 0.44) kcal/min, RMSE = 0.99 (SD = 0.62) METs, MAPE = 28.4% (SD = 11.8)
Existing lab, FL and RL models: EE overestimated during play sessions with low total EE and underestimated during sessions with high total EE
?
Steenbock et al. (2019) [88] n = 41
Age = 4.8 ± 0.8 (3 to 6) years
Sex = 46% girls
Absolute (kJ/min) and relative (kg/min/J) EE
Free-living (in- and outdoor semi-structured activities at school)
Indirect calorimetry using breath-by-breath VO2 and VCO2 to determine EE A Right hip AG-LM- HA/VA/DA-30 RMSE = 2.91 (SD = 0.95) kJ/min, RMSE = 124.91 (SD = 30.58) J/min/kg, RMSE = 1.70 (SD = 0.39) METs ? Very low
AG-MLM- HA/VA/DA-30 RMSE = 2.91 (SD = 0.99) kJ/min, RMSE = 125.40 (SD = 31.50) J/min/kg, RMSE = 1.70 (SD = 0.40) METs ?
AG-RF- HA/VA/DA-30 RMSE = 2.74 (SD = 0.96) kJ/min, RMSE = 115.56 (SD = 27.35) J/min/kg, RMSE = 1.56 (SD = 0.36) METs ?
AG-ANN- HA/VA/DA-30 RMSE = 2.86 (SD = 0.95) kJ/min, RMSE = 121.51 (SD = 29.57) J/min/kg, RMSE = 1.66 (SD = 0.39) METs ?
GA-LM- HA/VA/DA-30 RMSE = 2.89 (SD = 0.95) kJ/min, RMSE = 123.79 (SD = 31.28) J/min/kg, RMSE = 1.68 (SD = 0.39) METs ?
GA-MLM- HA/VA/DA-30 RMSE = 2.90 (SD = 0.98) kJ/min, RMSE = 124.70 (SD = 32.95) J/min/kg, RMSE = 1.69 (SD = 0.42) METs ?
GA-RF- HA/VA/DA-30 RMSE = 2.73 (SD = 1.00) kJ/min, RMSE = 112.57 (SD = 28.83) J/min/kg, RMSE = 1.53 (SD = 0.38) METs ?
GA-ANN- HA/VA/DA-30 RMSE = 2.83 (SD = 0.98) kJ/min, RMSE = 118.04 (SD = 29.80) J/min/kg, RMSE = 1.66 (SD = 0.43) METs ?
Left hip AG-LM- HA/VA/DA-30 RMSE = 2.84 (SD = 0.94) kJ/min, RMSE = 123.64 (SD = 31.82) J/min/kg, RMSE = 1.67 (SD = 0.41) METs ?
AG-MLM- HA/VA/DA-30 RMSE = 2.81 (SD = 0.99) kJ/min, RMSE = 123.58 (SD = 31.94) J/min/kg, RMSE = 1.67 (SD = 0.41) METs ?
AG-RF- HA/VA/DA-30 RMSE = 2.60 (SD = 0.97) kJ/min, RMSE = 112.32 (SD = 28.40) J/min/kg, RMSE = 1.52 (SD = 0.38) METs ?
AG-ANN- HA/VA/DA-30 RMSE = 2.78 (SD = 1.01) kJ/min, RMSE = 116.89 (SD = 28.41) J/min/kg, RMSE = 1.63 (SD = 0.42) METs ?
Right wrist GA-LM- HA/VA/DA-30 RMSE = 2.85 (SD = 0.83) kJ/min, RMSE = 125.60 (SD = 28.52) J/min/kg, RMSE = 1.69 (SD = 0.38) METs ?
GA-MLM- HA/VA/DA-30 RMSE = 2.83 (SD = 0.87) kJ/min, RMSE = 125.42 (SD = 28.74) J/min/kg, RMSE = 1.69 (SD = 0.38) METs ?
GA-RF- HA/VA/DA-30 RMSE = 2.56 (SD = 0.83) kJ/min, RMSE = 109.34 (SD = 26.98) J/min/kg, RMSE = 1.48 (SD = 0.37) METs ?
GA-ANN- HA/VA/DA-30 RMSE = 2.74 (SD = 0.88) kJ/min, RMSE = 115.91 (SD = 28.55) J/min/kg, RMSE = 1.62 (SD = 0.40) METs ?
Non-dominant wrist GA-LM- HA/VA/DA-30 RMSE = 2.83 (SD = 0.86) kJ/min, RMSE = 125.21 (SD = 27.38) J/min/kg, RMSE = 1.70 (SD = 0.37) METs ?
GA-MLM- HA/VA/DA-30 RMSE = 2.83 (SD = 0.89) kJ/min,
RMSE = 124.91 (SD = 27.58) J/min/kg, RMSE = 1.69 (SD = 0.38) METs
?
GA-RF- HA/VA/DA-30 RMSE = 2.56 (SD = 0.83) kJ/min, RMSE = 108.64 (SD = 26.33) J/min/kg, RMSE = 1.47 (SD = 0.36) METs ?
GA-ANN- HA/VA/DA-30 RMSE = 2.72 (SD = 0.91) kJ/min, RMSE = 114.01 (SD = 27.88) J/min/kg, RMSE = 1.61 (SD = 0.44) METs ?
Right thigh aP-LM- HA/VA/DA-30 RMSE = 2.94 (SD = 0.91) kJ/min, RMSE = 126.22 (SD = 29.46) J/min/kg, RMSE = 1.69 (SD = 0.38) METs ?
aP-MLM- HA/VA/DA-30 RMSE = 2.92 (SD = 0.95) kJ/min, RMSE = 125.96 (SD = 29.86) J/min/kg, RMSE = 1.69 (SD = 0.39) METs ?
aP-RF- HA/VA/DA-30 RMSE = 2.76 (SD = 0.94) kJ/min, RMSE = 115.61 (SD = 27.93) J/min/kg, RMSE = 1.56 (SD = 0.38) METs ?
aP-ANN- HA/VA/DA-30 RMSE = 3.08 (SD = 1.00) kJ/min, RMSE = 125.98 (SD = 31.50) J/min/kg, RMSE = 1.81 (SD = 0.40) METs ?
Butte et al. (2014) [75] n = 50
Age = 4.5 ± 0.8 (3 to 5) years
Sex = 50.0% girls
cpm
Laboratory (structured activities)
Indirect calorimetry using breath-by-breath VO2 and VCO2 to determine minute-by-minute EE I Right hip AG-CSTS-VA/HA/DA/VM-60 MAPE = -0.2 ± 6.7%, MAE = -0.005 ± 0.07 kcal/min, RMSE = 0.07 kcal/min, R2 = .88  +  Very low
AG-MARS-VA/HA/DA/VM-60 MAPE = 1.1 ± 6.6%, MAE = 0.009 ± 0.07 kcal/min, RMSE = 0.07 kcal/min ?
n = 105
Age = 4.6 ± 0.9 years
Sex = 66.7% girls
Activity intensity: SB, LPA, MPA, and VPA
Free-living
Indirect calorimetry using DLW to determine AEE I AG-B-VM-60 accuracy = 70%
SB: Sp = 83%
LPA: Sp = 64%
MPA: Sp = 35%
VPA: Sp = 38%
MVPA collapsed: accuracy = 74%
-
(SB + ; PA -)
AG-B-HA-60 accuracy = 68%
SB: Sp = 82%
LPA: Sp = 58%
MPA: Sp = 37%
VPA: Sp = 27%
MVPA collapsed: accuracy = 71%
-
(SB + ; PA -)
Chest Ah-B-HA-60 accuracy = 70%; Sp:
SB: Sp = 81%
LPA: Sp = 64%
MPA: Sp = 48%
VPA: Sp = 39%
MVPA collapsed: accuracy = 74%
-
(SB + ; PA -)
Convergent validity Sirard et al. (2005) [81] n = 269 (n = 69 3-, n = 125 4-, n = 75 5-year-olds)
Sex = 58% 3-, 52% 4-, 52% 5-year-old girls
Activity intensity: SB, LPA, MPA and VPA
Free-living (at preschool)
DVO using CARS to score activity type and derive activity intensity (SB: stationary, motionless or with trunk/limb movements; LPA: slow/easy movement; MPA: moderate movement; VPA: fast movement) by a single rater (beginning and end of video by four additional raters) VG Right hip AG-Si-VA-15 Total counts: rp = .58***
SB: rp = .70***, Se (94.4 to 100%), Sp (91.7 to 100%)
LPA: rp = .59***
VPA: rp = .61***, Se (95.8 to 100%), Sp (80.0 to 83.3%)
MPA: rp = .50***, Se (86.7 to 94.4%), Sp (66.7 to 100%)
MVPA: rp = .46***
+ 
(SB + ; PA +)
High
(level 1)
Peirera et al. (2020) [74] n = 60
Age = 2.7 ± 0.4 (1.8 to 3.5) years
Sex = 50% girls
% time spent in SB
Free-living (at childcare)
aP-PRE-uni-n.r. at right thigh
(SB: sitting/lying)
VG Hip   All outside the ± 10% equivalent interval of (-4.05 to 4.05)   Moderate
(level 2)
AG-C-VA-5 bias = -5.11, 90% CI (-7.90 to -2.32) ?
AG-Ev-VA-15 bias = 8.15, 90% CI (5.44 to 10.86) ?
AG-T12-VA-15 bias = 4.46, 90% CI (1.65 to 7.28) ?
AG-K-VA-15 bias = -7.95, 90% CI (-11.01 to -4.89) ?
AG-Pa-VA-15 bias = 6.17, 90% CI (3.40 to 8.93) ?
AG-Pa2-VA-15 bias = -9.26, 90% CI (-12.32 to -6.21) ?
AG-Re-VA-60 bias = -40.52, 90% CI (-43.82 to -37.22) ?
AG-Si-VA-15 bias = -15.07, 90% CI (-18.16 to -11.97) ?
AG-vC-VA-15 bias = -18.71, 90% CI (-21.89 to -15.54) ?
Dobell et al. (2019) [58] n = 62
Age = 3.5 ± 0.5 (3 to 5) years
Sex = 45.5% girls
Activity intensity: SB, LPA, and MVPA
Free-living (free-play in nursery)
DO using OSRAC-P (modified) to score activity type and derive activity intensity (SB: stationary/ motionless; LPA: stationary torso with limb and slow/easy movement; MVPA: moderate and fast movement) by two independent raters VG Non-dominant wrist AG-cts-VM-5/10/15/30 rsp = .34, R2 = .11* (5 s) to rsp = .48, R2 = .23* (30 s)
SB: AUC-ROC (.71 to .78), Se (.64 to .68), Sp (.64 to .78)
LPA: AUC-ROC (.55 to .56), Se (.74 to .80), Sp (.33 to .43)
MVPA: AUC-ROC (.67 to .78), Se (.63 to .80), Sp (.80 to .82)
-
(SB -;
PA -)
Moderate
(level 2)
Right hip AG-cts-VA-5/10/15/30 rsp = .38, R2 = .16* (5 s) to rsp = .52, R2 = .27* (10 s)
SB: AUC-ROC (.69 to .77), Se (.61 to .72), Sp (.69 to .72)
LPA: AUC-ROC (.53 to .54), Se (.59 to .76), Sp (.39 to .50)
MVPA: AUC-ROC (.71 to .81), Se (.55 to .70), Sp (.77 to .79)
-
(SB -;
PA -)
AG-cts-VM-5/10/15/30 rsp = .40, R2 = .16* (5 s) to rsp = .55, R2 = .30* (30 s)
SB: AUC-ROC (.73 to .79), Se (.69 to .72), Sp (.67 to .75)
LPA: AUC-ROC = .55, Se (.78 to .82), Sp (.35 to .36)
MVPA: AUC-ROC (.72 to .82), Se (.68 to .78), Sp (.62 to .79)
-
(SB -; PA -)
Ahmadi, Pavey, & Trost (2020) [85] n = 31
Age = 4.0 ± 0.9 (3 to 5) years
Sex = 29.0% girls
Activity intensity: SB, LPA, MVPA, run and walk
Free-living (20 min active play session)
DVO using customized scheme to score activity type and derive activity intensity (SB: sitting/lying down, and stationary, motionless; LPA: standing, stationary, movement of the limbs/trunk, easy translocation, MPA: moderate and fast translocation; walk: translocation, medium speed; run: translocation (very) fast) VG Right hip AG-RF-VM-1 Time and frequency features: F-score = 70.6%, + temporal features: F-score = 75.8%
SB: F-score = 73.3%, + temporal: F-score = 80.0%
LPA: F-score = 76.0%, + temporal: F-score = 80.4%
MVPA: F-score = 61.0%, + temporal: F-score = 65.2%
Walk: F-score = 55.7%, + temporal: F-score = 62.9%
Run: F-score = 53.1%, + temporal: F-score = 67.2%
- Low
(level 1)
AG-RF-VM-5 Time and frequency features: F-score = 79.5%, + temporal features: F-score = 82.2%
SB: F-score = 80.6%, + temporal: F-score = 85.7%
LPA: F-score = 82.7%, + temporal: F-score = 85.6%
MVPA: F-score = 75.5%, + temporal: F-score = 76.3%
Walk: F-score = 69.8%, + temporal: F-score = 68.8%
Run: F-score = 71.2%, + temporal: F-score = 72.8%
- (incl. temp
oral ±)
AG-RF-VM-10 Time and frequency features: F-score = 83.1%, + temporal features: F-score = 85.3%
SB: F-score = 82.3%, + temporal: F-score = 87.6%
LPA: F-score = 85.7%, + temporal: F-score = 87.9%
MVPA: F-score = 77.6%, + temporal: F-score = 76.4%
Walk: F-score = 80.7%, + temporal: F-score = 81.0%
Run: F-score = 74.4%, + temporal: F-score = 73.6%
 ± 
  AG-RF-VM-15 Time and frequency features: F-score = 84.0%, + temporal features: F-score = 85.9%
SB: F-score = 85.0%, + temporal: F-score = 87.7%
LPA: F-score = 86.8%, + temporal: F-score = 88.2%
MVPA: F-score = 75.3%, + temporal: F-score = 78.5%
Walk: F-score = 78.4%, + temporal: F-score = 79.4%
Run: F-score = 80.0%, + temporal: F-score = 82.6%
± 
Non-dominant wrist AG-RF-VM-1 Time and frequency features: F-score = 62.6%, + temporal features: F-score = 68.8%
SB: F-score = 63.3%, + temporal: F-score = 69.2%
LPA: F-score = 71.4%, + temporal: F-score = 75.5%
MVPA: F-score = 45.7%, + temporal: F-score = 57.5%
Walk: F-score = 45.9%, + temporal: F-score = 54.8%
Run: F-score = 55.1%, + temporal: F-score = 61.6%
-
AG-RF-VM-5 Time and frequency features: F-score = 70.8%, + temporal features: F-score = 75.5%
SB: F-score = 69.3%, + temporal: F-score = 78.%
LPA: F-score = 76.5%, + temporal: F-score = 80.1%
MVPA: F-score = 60.9%, + temporal: F-score = 66.9%
Walk: F-score = 60.7%, + temporal: F-score = 60.3%
Run: F-score = 68.5%, + temporal: F-score = 68.8%
-
AG-RF-VM-10 Time and frequency features: F-score = 74.5%, + temporal features: F-score = 80.0%
SB: F-score = 73.7%, + temporal: F-score = 82.4%
LPA: F-score = 79.1%, + temporal: F-score = 82.9%
MVPA: F-score = 62.0%, + temporal: F-score = 70.3%
Walk: F-score = 68.8%, + temporal: F-score = 70.8%
Run: F-score = 73.4%, + temporal: F-score = 71.5%
- (incl. temp
oral ±)
AG-RF-VM-15 Time and frequency features: F-score = 77.3%, + temporal features: F-score = 80.6%
SB: F-score = 78.2%, + temporal: F-score = 83.3%
LPA: F-score = 81.2%, + temporal: F-score = 83.7%
MVPA: F-score = 62.1%, + temporal: F-score = 70.7%
Walk: F-score = 70.5%, + temporal: F-score = 69.0%
Run: F-score = 82.4%, + temporal: F-score = 82.4%
- (incl. temp
oral ±)
Right hip + Non-dominant wrist AG-RF-VM-1 Time and frequency features: F-score = 72.5%, + temporal features: F-score = 76.8%
SB: F-score = 75.8%, + temporal: F-score = 80.8%
LPA: F-score = 77.7%, + temporal: F-score = 81.0%
MVPA: F-score = 62.1%, + temporal: F-score = 66.7%
Walk: F-score = 58.4%, + temporal: F-score = 64.9%
Run: F-score = 64.5%, + temporal: F-score = 67.9%
-
AG-RF-VM-5 Time and frequency features: F-score = 79.6%, + temporal features: F-score = 82.1%
SB: F-score = 81.6%, + temporal: F-score = 86.0%
LPA: F-score = 82.8%, + temporal: F-score = 85.6%
MVPA: F-score = 74.3%, + temporal: F-score = 74.5%
Walk: F-score = 70.1%, + temporal: F-score = 69.5%
Run: F-score = 71.7%, + temporal: F-score = 72.5%
- (incl. temp
oral ±)
AG-RF-VM-10 Time and frequency features: F-score = 83.2%, + temporal features: F-score = 85.3%
SB: F-score = 83.9%, + temporal: F-score = 87.5%
LPA: F-score = 85.8%, + temporal: F-score = 87.9%
MVPA: F-score = 75.6%, + temporal: F-score = 77.2%
Walk: F-score = 79.5%, + temporal: F-score = 80.7%
Run: F-score = 74.9%, + temporal: F-score = 73.0%
- (incl. temp
oral ±)
AG-RF-VM-15 Time and frequency features: F-score = 84.3%, + temporal features: F-score = 86.4%
SB: F-score = 85.5%, + temporal: F-score = 88.7%
LPA: F-score = 86.4%, + temporal: F-score = 88.4%
MVPA: F-score = 77.1%, + temporal: F-score = 78.0%
Walk: F-score = 78.1%, + temporal: F-score = 79.8%
Run: F-score = 86.6%, + temporal: F-score = 86.8%
± 
Hislop et al. (2012) [70] n = 31
Age = 4.4 ± 0.8 years
Sex = 51.6% girls
cpm and % time spent in MVPA
Free-living (unstructured outdoor free play during preschool)
DVO using CARS (modified) to score activity type and derive activity intensity (MVPA: moderate/fast translocation) by one rater VG Right hip AG-cts-VA-15 rsp = .39* - Low
(level 1)
RT-cts-VM-15 rsp = .56** -
AG-Si-VA-15 MD = 0.8 min* ?
RT-Swr-VM-15 MD = -12.2 min* ?
RT-Slj-VM-15 MD = 0 min ?
Hislop et al. (2012) [61] n = 31
Age = 4.4 ± 0.8 years
Sex = 51.6% girls
Activity intensity: SB, LPA, and MVPA
Free-living (unstructured outdoor free play during preschool)
DVO using CARS to score activity type and derive activity intensity (SB: stationary, motionless or with movement of limbs/trunk; LPA: easy translocation; MVPA: moderate/fast translocation) by one rater VG Right hip AG-Ev-VA-15 SB: MD = 13.2, LoA (-2.2 to 28.6)** ? Low
(level 1)
AG-Si-VA-15 SB: MD = -7.2, LoA (-20.2 to 5.7)**; LPA: MD = 6.3, LoA (-6.0 to 18.6)**
MVPA: MD = 0.8, LoA (-6.2 to 7.8)
?
AG-Puy-VA-15 SB: MD = 0.4, LoA (-13.25 to 13.99); LPA: MD = -2.2, LoA (-15.0 to 10.5)
MVPA: MD = 1.7, LoA (-8.0 to 11.5)
?
AG-vC-VA-15 SB: MD = -8.2, LoA (-22.7 to 6.2)**; LPA: MD = 11.4, LoA (-1.4 to 24.2)**
MVPA: MD = -3.3, LoA (-11.8 to 5.3)**
?
AG-Pa-VA-15 MVPA: MD = -8.7, LoA (-19.9 to 2.5)** ?
AG-Re-VA-15 SB: MD = -3.6, LoA (-17.6 to 10.4)** ?
Hislop et al. (2016) [76] n = 32
Age = 4.20 ± 0.5 (3 to 5) years
Sex = 34.4% girls
Activity intensity: SB, LPA, MVPA and total PA
Free-living
(free-play in nursery)
DVO using CARS (modified) to score activity type and derive activity intensity (SB: stationary, motionless; LPA: stationary, movement of the trunk/limbs and easy translocation; MVPA: moderate and fast translocation) VG Non-dominant wrist AG-J13-VM-5 Total PA: MD = 1.1, LoA (-9.9 to 12.1)
Locomotion CARS vs. high intensity: MD = -2.0, LoA (-9.4 to 5.4)
No clear agreement MVPA: MD = -9.3, LoA (-20.0 to 1.5); Locomotion vs. high wrist estimate: MD = 3.4, LoA (-3.1 to 10.0); Standing vs. low wrist estimate: MD = 2.8, LoA (-10.3 to 16.4)
? Low
(level 1)
Waist AG-Ev-VA-5 No clear agreement: Total PA: MD = 6.3, LoA (-8.8 to 21.4); MVPA: MD = -3.8, LoA (-10.7 to 3.1); Locomotion: MD = 3.4, LoA (-3.1 to 10.0); Standing: MD = 2.8, LoA (-10.3 to 16.4) ?
Martin et al. (2011) [72] n = 23
Age = 4.5 ± 0.7 years
Sex = 60.9% girls
SB and total PA
Free-living
aP-PRE-uni-n.r. at right thigh (SB: sitting/lying, and quiet standing) VG Right hip AG-Re-VA-60 rsp = .66***, MD = -2.1 ± 4.6%, LoA (-11.4 to 7.2%)*
Corrected (counts increased by 9%): rsp = .68***, MD = -4.3 ± 4.8%, LoA (-14.0 to 5.4%)***
+  Low
(level 2)
van Cauwenberghe et al. (2012) [64] n = 49
Age = 4.0 ± 0.5 (3 to 4) years
Sex = 55.1% girls
SB and total PA
Free-living (at childcare)
aP-PRE-uni-n.r. at right thigh
(SB: sitting/lying)
VG Right hip Ac-Ev-omni-15 accuracy = 73.0%, κ = .46, 95% CI (.45 to .47)
MD = 1.3 min; 0.1%, LoA (-59.4 to 61.9); 39.2%
- Low
(level 2)
Kahan, Nicaise, & Reuben (2013) [71] n = 12
Age = 4.7 ± 0.3 years
Sex = 75.0% girls
Activity intensity: SB and MVPA
Free-living (unstructured outdoor free play)
DO using OSRAC-P to score activity type and derive activity intensity (SB: stationary, motionless or with limb/trunk movement, light activity; MVPA: moderate/ vigorous activity) scored by one rater VG Right hip AG-Si-VA-5 accuracy = 81.2, 95% CI (79.2 to 83.1), κ = .48
SB: Se = 81.4%, Sp = 50.7%, bias = -8.1%
MVPA: Se = 35.8%, Sp = 87.6%, bias = 4.7%
± 
(SB ± ; PA ±)
Low
(level 2)
AG-vC-VA-5 accuracy = 78.5, 95% CI (76.4 to 80.3), κ = .48
SB: Se = 80.0%, Sp = 52.4%, bias = -8.6%
MVPA: Se = 53.9%, Sp = 79.1%, bias = -5.3%
-
(SB ± ; PA -)
AG-Ev-VA-5 accuracy = 79.9, 95% CI (77.5 to 82.1), κ = .58
SB: Se = 75.2%, Sp = 66.4%, bias = 36.3%***
MVPA: Se = 54.8%, Sp = 78.9%, bias = -5.9%
-
(SB -; PA -)
AG-Pa-VA-5 accuracy = 77.2, 95% CI (74.9 to 79.4), κ = .53
SB: Se = 50.0%, Sp = 58.7%, bias = 33.2%***
MVPA: Se = 62.7%, Sp = 72.2%, bias = -15.8%***
-
(SB -; PA -)
Alhassan et al. (2017) [56] n = 33
Age = 4.4 ± 0.8 (2.9 to 5) years
Sex = 36.4% girls
Counts
Free-living (at preschool)
Aw-cts-uni-15 at non-dominant wrist VG Waist AG-cts-VA-15 rsp = .41*** - Low
(level 2)
Activity intensity: SB, LPA, MPA, and VPA
Free-living
(at preschool)
DO using OSRAC-P (modified) to score activity type and derive activity intensity (SB: stationary, motionless or with trunk/limb movement; LPA: easy translocation; MPA: moderate translocation; VPA: vigorous translocation) by one rater A
VG (for MVPA)
AG-cts-VA-15 rsp = .47*** - Low (MVPA)
Very low (rest)
(level 2)
AG-Si-VA-15 accuracy = 50.3%, κ = .13, κw = .19
SB: accuracy = 62.7%, κ = .26, Se = 87.4%, Sp = 38.2%
LPA: accuracy = 65.0%, κ = .01, Se = 18.5%, Sp = 82.2%
MPA: accuracy = 86.3%, κ = .06, Se = 10.0%, Sp = 95.1%
VPA: accuracy = 86.5%, κ = .06, Se = 5.3%, Sp = 98.5%
MVPA: accuracy = 76.9%, κ = .15, Se = 16.2%, Sp = 95.2%
-
(SB -; PA -)
AG-Pa-VA-15 accuracy = 44.1%, κ = .21, κw = .32
SB: accuracy = 61.5%, κ = 0.40, Se = 60.1%, Sp = 79.7%
LPA: accuracy = 60.7%, κ = .13, Se = 51.1%, Sp = 64.3%
MPA: accuracy = 80.3%, κ = .09, Se = 24.0%, Sp = 86.8%
VPA: accuracy = 86.1%, κ = .19, Se = 18.0%, Sp = 96.2%
MVPA: accuracy = 75.3%, κ = .27, Se = 39.8%, Sp = 86.0%
-
(SB -; PA -)
Non-dominant wrist Aw-cts-uni-15 rsp = .47*** -
Aw-Ek-uni-15 accuracy = 50.3%, κ = .13, κw = .19
SB: accuracy = 65.47%, κ = 0.31, Se = 41.5%, Sp = 89.1%
LPA: accuracy = 53.8%, κ = .10, Se = 62.3%, Sp = 50.7%
MPA: accuracy = 82.8%, κ = .11, Se = 22.0%, Sp = 89.7%
VPA: accuracy = 86.3%, κ = .31, Se = 33.9%, Sp = 94.0%
MVPA: accuracy = 77.5%, κ = .35, Se = 46.9%, Sp = 86.8%
-
(SB -; PA -)
Ahmadi et al. (2020) [83] n = 31
Age = 4.0 ± 0.9 (3 to 5) years
Sex = 29% girls
Activity intensity: SB, LPA, MVPA, run, and walk
Free-living (play session)
DVO using customized scheme to score activity type and derive activity intensity (SB: sitting/lying down, and stationary, motionless; LPA: standing, stationary, movement of the limbs/trunk, easy translocation, MPA: moderate and fast translocation; walk: translocation, medium speed; run: translocation (very) fast) A Non-dominant wrist AG-RF-VM-15 accuracy = 59.1%, 95% CI (57.1 to 61.1%), κ = .37
SB: accuracy = 76.8%
LPA: accuracy = 57.5%
MVPA: accuracy = 83.6%
Walk: accuracy = 14.8%
Run: accuracy = 68.7%
Excl. multiple activities: walk accuracy (14.8 to 43.8%) and run accuracy (68.7 to 100%)
-
(SB -; PA -)
Very low
(level 1)
AG-SVM-VM-15 accuracy = 59.3%, 95% CI (57.3 to 61.3%), κ = .37
SB: accuracy = 76.8%
LPA: accuracy = 58.3%
MVPA: accuracy = 82.6%
Walk: accuracy = 12.3%
Run: accuracy = 70.1%
Excl. multiple activities: walk accuracy (12.3 to 45.8%) and run accuracy (70.1 to 100%)
-
(SB -; PA -)
Right hip AG-RF-VM-15 accuracy = 69.4%, 95% CI (67.4 to 71.2%), κ = .48
SB: accuracy = 71.7%
LPA: accuracy = 79.3%
MVPA: accuracy = 71.3%
Walk: accuracy = 10.4%
Run: accuracy = 74.5%
Excl. multiple activities: walk accuracy (10.4 to 29.2%) and run accuracy (74.5 to 100%)
-
(SB -; PA -)
AG-SVM-VM-15 accuracy = 66.4%, 95% CI (64.4 to 68.3%), κ = .45
SB: accuracy = 77.7%
LPA: accuracy = 71.1%
MVPA: accuracy = 73.0%
Walk: accuracy = 8.9%
Run: accuracy = 65.8%
Excl. multiple activities: walk accuracy (8.9 to 33.3%) and run accuracy (65.8 to 100%)
-
(SB -; PA -)
Trost et al. (2018) [89] n = 11
Age = 4.8 ± 0.87 (3 to 6) years
Sex = 55% girls
Activity intensity: SB, LPA, MVPA, run, and walk
Laboratory (semi-standardized activities
DO to score activity type and determine activity intensity based on preschool-specific MET thresholds (SB < 1.5 METs, LPA 1.5–2.7, MPA 2.8–3.4 VPA ≥ 3.5) A Non-dominant wrist AG-SVM-VM-15 accuracy = 80.4%, 95% CI (78.9 to 81.9%), κw = .73, CV κw = .78
SB: accuracy = 90.0%
LPA: accuracy = 73.7%
MVPA: accuracy = 78.6%
Walk: accuracy = 70.6%
Run: accuracy = 71.2%
+ 
(SB + ; PA -)
Very low
(level 2)
AG-RF-VM-15 accuracy = 78.1%, 95% CI (76.6 to 79.9%), κw = .70, CV κw = .76
SB: accuracy = 89.1%
LPA: accuracy = 68.7%
MVPA: accuracy = 79.1%
Walk: accuracy = 61.1%
Run: accuracy = 68.8%
± 
(SB + ; PA -)
AG-T18-VA-15 κw = .64 -
AG-T18-VM-15 κw = .49 -
Right hip AG-SVM-VM-15 accuracy = 81.3%, 95% CI (79.9 to 82.8%), κw = .74, CV κw = .82
SB: accuracy = 94.0%
LPA: accuracy = 76.8%
MVPA: accuracy = 76.0%
Walk: accuracy = 63.5%
Run: accuracy = 69.8%
+ 
(SB + ; PA -)
AG-RF-VM-15 accuracy = 80.2%, 95% CI (79.9 to 81.6%), κw = .72, CV κw = .81
SB: accuracy = 93.2%
LPA: accuracy = 72.1%
MVPA: accuracy = 76.7%
Walk: accuracy = 61.1%
Run: accuracy = 73.0%
± 
(SB + ; PA -)
AG-T18-VA-15 κw = .65 -
AG-T18-VM-15 κw = .61 -
Non-dominant wrist + hip AG-SVM-VM-15 accuracy = 85.2%, 95% CI (83.8 to 86.5%), κw = .80, CV κw = .84
SB: accuracy = 94.7%
LPA: accuracy = 80.2%
MVPA: accuracy = 81.7%
Walk: accuracy = 69.7%
Run: accuracy = 82.3%
+ 
(SB + ; PA +)
AG-RF-VM-15 accuracy = 81.8%, 95% CI (80.4 to 83.2%), κw = .75, CV κw = .81
SB: accuracy = 93.0%
LPA: accuracy = 73.9%
MVPA: accuracy = 80.7%
Walk: accuracy = 63.0%
Run: accuracy = 75.3%
 + 
(SB + ; PA +)
Finn & Specker (2000) [60] n = 40
Age = (3 to 4) years
Sex = 60% girls
Mean activity counts (3-min)
Free-living
DVO using CARS (3-min average) to score activity type and derive activity intensity (SB: stationary, motionless or with movement of limbs/trunk; LPA: easy translocation; MVPA: moderate/fast translocation) by two independent raters D Waist Aw-cts-uni-60 r = .03 to r = .92 (median = .74)
More active children higher association between CARS and counts: within-child correlation coefficients vs. CARS r = .37* or counts r = .31
PA: - Very low
(level 1)
Brønd et al. (2020) [86] n = 29
Age = 4.4 (3 to 6) years
Sex = n.r. % girls
Activity type: Sit, stand, run, walk, and bike
Free-living (structured activities at school)
DO to score activity type (sit: on floor/chair playing; stand: stand close to table playing; run: run together with instructor; walk: walks together with instructor; bike: on running bike) by one rater D Thigh AX-DT-HA/VA/DA-2 Sit: Se = 100%, Sp = 100%
Stand: Se = 100%, Sp = 99.8%
Bike (sitting bike): Se = 64.8%, Sp = 100%
Walk: Se = 82.6%, Sp = 98.1%
Run: Se = 92.4%, Sp = 95.0%
+  Very low
(level 2)
Zhao et al. (2013) [90] n = 69
Age = (3 to 5) years
Sex = 55% girls
Activity type: Sleep, rest, quiet play, low active play, moderately active play, and very active play
Laboratory (structured activities)
DO to score activity type (sleep: lying and sleeping; rest: watching tv, quiet play: sitting while coloring, puzzle, video game; low active play: stand and play; moderately active play: standing while ball toss, active video game, dance; very active play: running in place) D Right hip AG-SVM -VA/HA/DA/VM-60 accuracy = 75.3%
Sleep: accuracy = 91.44%
Rest: accuracy = 65.66%
Quiet play: accuracy = 74.07%
Low active play: accuracy = 68.49%
Moderately active play: accuracy = 93.73%
Very active play: accuracy = 98.73%
Misclassification highest for sleep (31.6%) and quiet play (33.0%)
±  Very low
(level 2)
  AG-MLR—VA/HA/DA/VM-60 accuracy (67.2 to 73.2%) -
Hagenbuchner et al. (2015) [87] n = 11
Age = 4.8 ± 0.87 (3 to 6) years
Sex = 55% girls
Activity intensity: SB, LPA, MVPA, run, and walk
Laboratory
(activity trials)
DO to score activity type and derive activity intensity (SB: watching tv, tablet/computer, reading, quiet play; LPA: clean up toys, standing art, treasure hunt; MVPA: bicycle riding, obstacle course, active game; walk: walk at self-selected pace; run: run at self-selected speed) D Right hip AG-MLP-VM-60 accuracy = 69.7%
SB: accuracy = 81.8%
LPA: accuracy = 78.8%
MVPA: accuracy = 63.6%
Walk: accuracy = 36.4%
Run: accuracy = 45.5%
-
(SB + ; PA -)
Very low
(level 2)
AG-SOM-VM-60 accuracy = 53.8%
SB: accuracy = 59.1%
LPA: accuracy = 57.6%
MVPA: accuracy = 66.7%
Walk: accuracy = 36.4%
Run: accuracy = 0.0%
-
(SB -; PA -)
AG-DLEN-VM-60 accuracy = 82.6%
SB: accuracy = 84.1%
LPA: accuracy = 90.9%
MVPA: accuracy = 78.8%
Walk: accuracy = 72.7%
Run: accuracy = 72.7%
± 
(SB + ; PA -)
de Bock et al. (2010) [68] n = 33
Age = 3.5 ± 0.8 (boys); 4.3 ± 1.1 (girls) years
Sex = 36.36% girls
Activity intensity: SB and MVPA
Free-living (at preschool)
DO using CARS to score activity type and derive activity intensity (SB: stationary, motionless or with trunk/limb movement; MVPA: easy, moderate, and fast translocation) by five independent raters D Chest Ah-dB-VA-15 Calibration
MVPA: AUC-ROC = .86
SB: AUC-ROC = .79
Cross-validation
SB: accuracy 67% (boys), 69% (girls); Sp = 52% (boys), 61% (girls); Se = 78% (boys), 75% (girls)
MVPA: accuracy = 84% (boys), 87% (girls)
-
(SB -;
PA +)
Very low
(level 2)
Fairweather et al. (1999) [59] n = 11
Age = 3.7 ± 0.5 years
Sex = 72.7% girls
Mean cpm
Free-living (exercise class)
DO using 1-min average CPAF score (ranging from 1 to 4) scored by one rater D Right hip CSA-cts-uni-60 r = .87***, rsp = .79** PA: +  Very low
(level 2)
n = 10
Age = 4.0 ± 0.4 years
Sex = 90% girls
CSA at left hip r = .92**, MD = 31 (SD = 64, 95% CI (1 to 61)*, rsp = .97**
Relative PA assessment largely unaffected by location
PA: + 
Klesges & Klesges (1987) [62] n = 28
Age = 33.9 ± 8.0 months
Sex = 43.3% girls
Total PA (all-day) and hourly PA
Free-living (unstructured activities)
DO using FATS to score behavior and intensity D Left hip CPAC-cts-uni-n.r r = .54*** (total PA); r = .62** to .95*** (hourly PA)
Behavior: lying rsp = -.37; sitting rsp = -.26; standing rsp = .60**; walking rsp = .50**; running rsp = .37
Intensity: minimal rsp = -.51**; moderate rsp = .51**; extreme rsp = .40*
Summary: correlations higher for older children (< 32.5 months rsp = .39; > 32.5 months rsp = .80***), girls (rsp = ,67*, boys rsp = .43), and obese children (rsp = .75*, normal weight rsp = .46*)
PA: ±  Very low
(level 2)
Klesges et al. (1985) [63] n = 30
Age = 47.7 ± 12.5 months
Sex = 40% girls
Total PA
Free living (free play)
DO using FATS to score behavior and intensity D Left hip CPAC-cts-uni-n.r Behavior: r = .38*, rsp = .28: lying r = .17, rsp = -.18; sitting r = -.31*, rsp = -.32*; crawling r = -.22, rsp = -.20; climbing r = -.06, rsp = .02; standing r = .35*, rsp = .37*; walking r = -.11, rsp = -.03; running r = .32*, rsp = .22
Intensity: r = .32*, rsp = .29: minimal r = -.27, rsp = -.23; moderate r = .05, rsp = .07; extreme r = .44**, rsp = .43**
Summary: r = .40**, rsp = .36*
PA: - Very low
(level 2)
Right hip LSI-cts-uni-n.r Behavior: r = .38*, rsp = .28: lying r = .17, rsp = -.18; sitting r = -.31*, rsp = -.32*; crawling r = -.22, rsp = -.20; climbing r = -.06, rsp = .02; standing r = .35*, rsp = .37*; walking r = -.11, rsp = -.03; running r = .32*, rsp = .22
Intensity: r = .32*, rsp = .29: minimal r = -.27, rsp = -.23; moderate r = .05, rsp = .07; extreme r = .44**, rsp = .43**
Summary: r = .40**, rsp = .36*
PA: -
Ettienne et al. (2016) [69] n = 30
Age = 3.5 ± 0.6 years
Sex = 43% girls
Activity intensity: SB, LPA, and MVPA
Free-living (during preschool)
DO using (modified) SOFIT to score activity type and derive activity intensity (SB: lying; LPA: sitting, standing; MVPA: walking, very active) D Non-dominant wrist Ac-Sc-omni-60 accuracy = 74%, κw = .17***
391 min classified as MVPA, whereas DO class was LPA; 406 min DO classified as MVPA and accelerometer as LPA
- Very low
(level 2)
Djafarian et al. (2013) [57] n = 42
Age = 4.1 ± 0.8 (3 to 5) years
Sex = 47.6% girls
cpm 1–10-min periods
Free-living (structured & unstructured)
DO using 1-min average CARS score D Non-dominant wrist Aw-cts-uni-300 r = .52** (n = 21 cross-validation sample)
Measured (DO) vs. predicted (log accelerometry) CARS score MD = 0.025 (± 0.38), CI 95% with 0.74 CARS unit
PA: - Very low
(level 2)
Davies et al. (2012) [82] n = 30
Age = 4.1 ± 0.5 (3.1 to 4.9) years
Sex = 60% girls
Activity type: Sit/lie, stand, and walk
Free-living (usual nursery setting)
DVO using to score activity type (walk, sit, stand, lie, other (crouching down, kneeling up, crawling) by one rater I Right thigh aP-PRE-uni-1 Time spent: sitting MD = -4.4%**, standing MD = 7.1%**
Sit/lie: Se = 86.7%, Sp = 97.1%, PPV = 96.3%
Stand: Se = 91.8%, Sp = 84.3%, PPV = 75.4%
Walk: Se = 80.3%, Sp = 95.9%, PPV = 78.4%
 +  Very low
(level 1)
Johansson et al. (2015) [77] Calibration
n = 26
Age = 26 ± 6.0 months
Sex = 38% girls
Cross-validation
n = 12
Age = 25 ± 5.6 (15 to 36) months
Sex = 50% girls
Activity intensity: SB, LPA, and MVPA
Free-living
(structured activities and free play at preschool)
DVO using CARS to score activity type and derive activity intensity (SB: stationary, motionless; LPA: stationary, movement of the limbs/trunk, easy translocation; MVPA: moderate and fast translocation) I Non-dominant wrist AG-J13-VM-5 Calibration
SB: AUC-ROC = .98, CI (.95 to 1.0)
MVPA: AUC-ROC = .90, CI (.81 to .98)
Cross-validation
SB: rsp = .91
LPA: rsp = .77*
MVPA: rsp = .69*
 + 
(SB + ; PA
 ±)
Very low
(level 1)
AG-J13-VA-5 Calibration
SB: AUC-ROC = .98, CI (.95 to 1.0)
MVPA: AUC-ROC = .88, CI (.78 to .97)
Cross-validation
SB: rsp = .91
LPA: rsp = .89*
MVPA: rsp = .77*
 + 
(SB + ; PA
 +)
Johansson et al. (2016) [78] n = 30
Age = 49.0 ± 3.7 months
Sex = 53% girls
Activity intensity: SB, LPA, and MVPA
Free-living
(structured indoor activities and free play at preschool)
DO using CARS to score activity type and derive activity intensity (SB: stationary, motionless; LPA: stationary, movement of the limbs/trunk, easy translocation; MVPA: moderate and fast translocation) by one rater I Left hip AG-J16-VM-5 κqw = .86, 95% CI (.85 to .87), accuracy = 75%
SB: Se = 100%, Sp = 60%, AUC-ROC = .95
MVPA: Se = 70%, Sp = 93%, AUC-ROC = .91
+ 
(SB ± ; PA ±)
Very low
(level 2)
AG-J16-VA-5 κqw = .76, 95% CI (.74 to .77), accuracy = 68%
SB: Se = 100%, Sp = 60%, AUC-ROC = .93
MVPA: Se = 70%, Sp = 100%, AUC-ROC = .95
± 
(SB ± ; PA ±)
Non-dominant wrist AG-J16-VM-5 κqw = .95, 95% CI (.94 to .96), accuracy = 82%
SB: Se = 100%, Sp = 60%, AUC-ROC = .99
MVPA: Se = 70%, Sp = 100%, AUC-ROC = .96
+ 
(SB ± ; PA ±)
AG-J16-VA-5 κqw = .92, 95% CI (.91 to .93), accuracy = 82%
SB: Se = 100%, Sp = 60%, AUC-ROC = .99
MVPA: Se = 70%, Sp = 100%, AUC-ROC = .95
+ 
(SB ± ; PA ±)
Reilly et al. (2003) [80] n = 52
Age = 3.5 ± 0.5 (3 to 4) years
Sex = 59.62% girls
Activity intensity: SB
Free-living
DO using CPAF to score activity type and derive activity intensity
(SB: stationary, motionless or with limb movement) by one rater
I Right hip CSA-Re-uni-60 Se = 83%, 95% CI (78 to 86%)
Sp = 82%, 95% CI (89 to 86%)
SB: +  Very low
(level 2)
Adolph et al. (2012) [51] n = 64
Age = 4.5 ± 0.8 (3 to 5) years
Sex = 42.2% girls
Activity intensity: SB, LPA, and MVPA
Laboratory
DO using CARS to score activity type and derive CARS levels (1 to 5) I Chest Ah-A-uni-60 accuracy = 72%, TPR = 82%
SB: accuracy = 75%
LPA: accuracy = 61%
MVPA: accuracy = 82%
-
(SB -; PA
 ±)
Very low
(level 2)
Right hip Ac-A-omni-60 accuracy = 71%, TPR = 69%
SB: accuracy = 77%
LPA: accuracy = 63%
MVPA: accuracy = 69%
-
(SB -; PA
-)
RT-A-VM-60 accuracy = 73%, TPR = 79%
SB: accuracy = 76%
LPA: accuracy = 65%
MVPA: accuracy = 79%
-
(SB -; PA
 ±)
Li et al. (2020) [79] n = 34
Age = 4.0 ± 0.5 (3 to 5) years
Sex = 58.8% girls
Activity intensity: SB, LPA, MPA, and VPA
Free-living (during preschool)
AG-B-VM-60 I Right hip AG-LiROC-VM-60 κ = .30, accuracy = 59.5%
SB: Se = 77.8%, Sp = 77.8%, FPR = 22.2%, FNR = 22.2%, κ = .54
LPA: Se = 26.8%, Sp = 87.5%, FPR = 12.5%, FNR = 73.2%, κ = .16
MPA: Se = 4.5%, Sp = 69.8%, FPR = 3.2%, FNR = 95.5%, κ = .02
VPA: = Se = 79.0%, Sp = 79.0%, FPR = 21.0%, FNR = 21.0%, κ = .11
MVPA collapsed: κ = .35, accuracy = 63.6%, Se = 78.9%, Sp = 78.9%, FPR = 20.1%, FNR = 21.1%, κ = .28
-
(SB -; PA -)
Very low
(level 2)
AG-LiOLR-VM-60 κ = .37, accuracy = 71.5%
SB: Se = 91.0%, Sp = 52.8%, FPR = 47.2%, FNR = 9.1%, κ = .47
LPA: Se = 43.3%, Sp = 86.5%, FPR = 56.7%, FNR = 13.5%, κ = .32
MPA: Se = 8.8%, Sp = 99.1%, FPR = 1.0%, FNR = 76.1%, κ = .12
VPA: Se = 23.9%, Sp = 99.0%, FPR = 1.0%, FNR = 76.1%, κ = .28
MVPA collapsed: κ = .39, accuracy = 72.4%, Se = 24.3%, Sp = 99.0%, FPR = 1.0%, FNR = 75.7%, κ = .33
-
(SB -; PA -)
AG-J13-VM-60 Non-dominant wrist AG-Li4k-VM-60 κ = .37, accuracy = 63.7%
SB: Se = 71.6%, Sp = 85.1%, FPR = 14.9%, FNR = 28.4%, κ = .53
LPA: Se = 50.9%, Sp = 76.8%, FPR = 23.2%, FNR = 49.1%, κ = .27
MPA: Se = 48.4%, Sp = 86.8%, FPR = 13.2%, FNR = 51.6%, κ = .19
VPA: Se = 33.9%, Sp = 98.2%, FPR = 1.8%, FNR = 66.1%, κ = .31
MVPA collapsed: κ = .40, accuracy = 65.6%, Se = 68.4%, Sp = 86.6%, FPR = 13.4%, FNR = 31.6%, κ = .35
-
(SB -; PA -)
Pagels, Boldemann, & Raustorp (2011) [73] n = 55
Age = 4.5 ± 0.6 (3.4 to 5.7) years
Sex = 49.1% girls
Activity intensity: LPA and MVPA
Free-living (at preschool)
Pedometer (Yamax SW200) to measure daily steps I Waist AG-Si-VA-15 r = .67***, R2adj = 45%: (3-year-olds: r = .37, 4-year-olds: r = .52**, 5-year-olds: r = .84***)
LPA: r = .76***, R2adj = 52%
MVPA: 3-year-olds (r = .19, R2adj = 4%); 4-to-5-year-olds (r = .50***, R2adj = 23%)
PA: - Very low
(level 3)
  1. Abbreviations: A Adolph’s cut-points (2012) [51], Ac Actical, ACT40 Actiware software with wake threshold value of 40, ACT80 Actiware software with wake threshold value of 80, AEE activity energy expenditure, AIC Akaike information criterion, AG ActiGraph, Ah Actiheart, ANN artificial neural network, aP activPAL, AS AlgoSmooth [94], AUC-ROC area under the receiver operating curve, Aw Actiwatch, AX Axivity, B Butte’s cut-points (2014) [75], C Costa’s cut-points (2014) [46], CARS children’s activity rating system, CCC concordance correlation coefficient, CI confidence interval, CPAC Caltrac personal activity computer, CPAF children’s physical activity form, cpm counts per minute, cts counts, CSA computer science and applications activity monitor, CSTS cross-sectional time series, CV cross-validation, DA diagonal axis (z-axis), dB de Bock’s cut-points (2010) [68], DLEN deep learning ensemble network, DLW doubly labelled water, DO direct observation, DT decision table, DVO direct video observation, EE energy expenditure, Ek Ekblom’s cut-points (2012) [95], Ev Evenson’s cut-points (2008) [91], FATS Fargo activity time-sampling survey, FNR false negative rate, FPR false positive rate, GA GENEActiv, HA horizontal axis (x-axis), ICC intraclass correlation coefficient, J13 Johansson’s cut-points (2015) [77], J16 Johansson’s cut-points (2016) [78], κ Kappa, K Kelly’s cut-points (2016) [96], κw weighted Kappa, κqw quadratic weighted Kappa, LiROC Li’s cut-points (2020) [79], LiORL Li’s cut-points (2020) [79], Li4k Li’s cut-points (2020) [79], LoA limits of agreement, LM linear model, LPA light physical activity, LSI LSI moving activity monitor, MAE mean absolute error, MAPE mean absolute percentage error, MARS multivariate adaptive regression splines, MD mean difference, METs metabolic equivalents, MLM mixed linear model, MLP multi-layer perceptron network, MLR multinomial logistic regression, MPA moderate physical activity, MVPA moderate-to-vigorous physical activity, n.r. not reported, NVP negative predictive value, omni omnidirectional, OSRAC-P observational system for recording physical activity in children preschool, Pa Pate’s cut-points (2006) [52], PA physical activity, Pa2 Pate’s cut-points (2006) [52], PAL physical activity level, Pf Pfeiffer’s cut-points (2006) [53], PPV positive predictive value, PRE activPAL professional research edition software, Puy Puyau’s cut-points (2002) [97], r correlation coefficient (unknown), Re Reilly’s cut-points (2003) [80], RF random forests, Ro Roscoe’s cut-points (2017) [67], rp correlation coefficient (Pearson), rsp correlation coefficient (Spearman rank), R2 R-squared value, R.2adj adjusted R-squared value, RMSE root mean square error, RT Research Tracker 3, SB sedentary behavior, Sc Schaefer’s cut-points (2014) [98], SD standard deviation, Se sensitivity, SE standard error, Si Sirard’s age-specific cut-points (2005) [81], Slj Sun’s cut point (2008) [99], SMR sleeping metabolic rate, SOFIT system for observing fitness instruction time, SOM self-organizing map, Sp specificity, SVM support vector machine, Swr Sun’s cut point (2008) [99], TD TracmorD, TEE total energy expenditure, tri mean of the three axes, TPR true positive rate, TST total sleep time, T12 Trost’s cut-points (2012) [44], T18 Trost’s cut-points (2018) [89], uni uniaxial (axis not specified), VA vertical axis (y-axis), vC van Cauwenberghe’s cut-points (2011) [45], VPA vigorous physical activity, VCO2 carbon dioxide, VM vector magnitude, VO2 oxygen consumption, WASO wake after sleep onset, 1 1 s epoch, 2 2 s epoch, 5 5 s epoch, 10 10 s epoch, 15 15 s epoch, 30 30 s epoch, 60 60 s epoch, 300 5 min epochs
  2. aAge presented as mean ± SD (range)
  3. bMethodological study quality based on newly developed checklist: VG very good, A adequate, D doubtful, I inadequate
  4. cDevice-based method described using code combinations of four elements resulting in the following format: brand-axis-approach-epoch length
  5. dStudy result rating based on COSMIN guideline: + sufficient, ± inconsistent, - insufficient, ? intermediate
  6. eQuality of evidence based on GRADE approach
  7. * p < .05
  8. ** p < .01
  9. *** p < .001