Table 2.

Results from agreement analysis and descriptive statistics for hind limb gait events. Displaying the three algorithms with the best (smallest) accuracy and precision, all based on features from DMC ^A/DMT ^B mounted IMUs ^C. All values are in ms. Values in bold have the best (smallest) accuracy and precision. The full table can be accessed online as supplementary material.

Hind Limb		Agreement				Descriptive Statistics
Gait Event	Feature	LLoA 1	ULoA 2	Bias	ICC 3	Median Error	SE Mean	SD
Hoof-on	Horizontal acceleration 4	−11.33	9.88	−0.73	0.9998	−5	2.45	26.60
	Vertical acceleration 5	−13.50	5.96	−3.77	0.9997	−20	2.28	24.79
	Vertical acceleration 6	−16.52	4.03	−6.25	0.9994	−30	2.38	25.89
Hoof-off	XYZ velocity magnitude	−3.69	15.98	6.15	0.9995	35	2.32	25.03
	Horizontal acceleration 6	−5.91	11.48	2.78	0.9998	20	2.05	22.04
	Horizontal displacement	−2.08	14.42	6.17	0.9995	35	1.95	21.00
Stance	Hoof-on: Vertical acceleration 4 Hoof-off: Horizontal acceleration 6	−54.39	54.66	0.13	0.9078	2	2.60	27.91
	Hoof-on: Vertical acceleration 4 Hoof-off: Horizontal displacement	−54.13	54.52	0.20	0.9021	0	2.58	27.70
	Hoof-on: Horizontal acceleration 4 Hoof-off: Horizontal displacement	−56.56	56.90	0.17	0.8971	−2	2.67	28.67

^ADMC: Distal MetaCarpus

^BDistal MetaTarsus

^CIMUs: Inertial Measurement Units

¹Lower limits of agreement

²Upper limits of agreement

³Intra Class Correlation

⁴Before rotation

⁵Before rotation, wavelet decomposed

⁶After rotation, before filtering.