Table 1 Cross-sectional properties of ankle submovement models and features

From: At-home wearables and machine learning sensitively capture disease progression in amyotrophic lateral sclerosis

Feature name	Stat.	SM dur. group	Dir. of motion group	Relationship with ALSFRS-R										Test re- test rel.	ALS vs Control		Right and left ankle agreement
				Total		Gross motor		Fine motor		Respiratory		Bulbar		Test re- test rel.	ALS vs Control		Right and left ankle agreement
				r	p-val	r	p-val	r	p-val	r	p-val	r	p-val	ICC	p-val	es	r	p-val
ALSFRS total prediction model				0.59	0	0.66	0	0.52	6.0E−293	0.26	2.0E−66	0.20	4.0E−41	0.88	2.0E−56	1.1	0.91	0
ALSFRS Gross Motor Prediction Model				0.54	0	0.77	0	0.40	3.0E−162	0.21	7.0E−43	0.11	7.0E−12	0.91	2.0E−78	1.4	0.95	0
Pairwise Model				0.56	0	0.75	0	0.43	2.0E−191	0.22	6.0E−49	0.16	4.0E−25	0.92	3.0E−73	1.3	0.94	0
SM Dist.	M	Long	PC1	0.36	9.0E−134	0.45	3.0E−212	0.23	7.0E−53	0.26	2.0E−68	0.09	9.0E−10	0.85	5.0E−48	1.0	0.90	0
	M	Long	PC2	0.48	2.0E−249	0.56	0	0.37	3.0E−144	0.30	2.0E−90	0.12	4.0E−15	0.90	8.0E−57	1.1	0.94	0
	M	Short	PC1	0.50	5.0E−269	0.56	0	0.46	3.0E−226	0.22	4.0E−47	0.13	7.0E−19	0.90	7.0E−42	0.9	0.92	0
	M	Short	PC2	0.54	0	0.60	0	0.48	9.0E−253	0.27	4.0E−70	0.15	2.0E−21	0.91	9.0E−42	0.9	0.93	0
	SD	Long	PC1	0.31	3.0E−96	0.40	7.0E−163	0.16	5.0E−27	0.23	7.0E−54	0.10	4.0E−10	0.71	3.0E−49	0.9	0.81	0
	SD	Long	PC2	0.39	5.0E−154	0.46	4.0E−228	0.26	4.0E−67	0.26	2.0E−66	0.12	7.0E−15	0.75	2.0E−59	1.1	0.86	0
	SD	Short	PC1	0.52	2.0E−305	0.63	0	0.46	5.0E−227	0.23	6.0E−51	0.14	7.0E−19	0.90	8.0E−64	1.2	0.91	0
	SD	Short	PC2	0.54	0	0.66	0	0.47	1.0E−231	0.24	6.0E−59	0.14	4.0E−21	0.88	6.0E−71	1.3	0.93	0
SM Vel.	M	Long	PC1	0.49	2.0E−264	0.56	0	0.37	3.0E−142	0.30	2.0E−90	0.16	5.0E−26	0.88	7.0E−56	1.1	0.92	0
	M	Long	PC2	0.57	0	0.65	0	0.47	8.0E−234	0.31	2.0E−95	0.16	3.0E−26	0.92	4.0E−63	1.2	0.96	0
	M	Short	PC1	0.50	7.0E−277	0.59	0	0.46	2.0E−225	0.21	2.0E−42	0.14	3.0E−19	0.91	2.0E−38	0.8	0.92	0
	M	Short	PC2	0.54	0	0.63	0	0.48	5.0E−248	0.25	8.0E−65	0.15	6.0E−22	0.91	3.0E−38	0.9	0.93	0
	SD	Long	PC1	0.37	7.0E−137	0.44	2.0E−202	0.22	5.0E−50	0.23	2.0E−54	0.14	1.0E−20	0.71	4.0E−55	1.0	0.83	0
	SD	Long	PC2	0.49	2.0E−262	0.59	0	0.36	2.0E−132	0.26	6.0E−67	0.17	3.0E−28	0.80	3.0E−77	1.3	0.91	0
	SD	Short	PC1	0.52	8.0E−294	0.63	0	0.45	7.0E−219	0.21	3.0E−45	0.13	3.0E−17	0.91	8.0E−52	1.1	0.91	0
	SD	Short	PC2	0.55	0	0.68	0	0.47	6.0E−237	0.24	5.0E−57	0.14	3.0E−20	0.90	2.0E−59	1.1	0.92	0
SM Accel.	M	Long	PC1	0.58	0	0.63	0	0.51	2.0E−285	0.26	4.0E−69	0.20	3.0E−37	0.93	5.0E−48	1.0	0.95	0
	M	Long	PC2	0.58	0	0.66	0	0.51	2.0E−284	0.27	1.0E−73	0.17	3.0E−30	0.93	2.0E−50	1.0	0.97	0
	M	Short	PC1	0.49	2.0E−262	0.59	0	0.44	3.0E−209	0.19	7.0E−38	0.13	7.0E−18	0.92	5.0E−37	0.8	0.91	0
	M	Short	PC2	0.53	0	0.63	0	0.47	2.0E−233	0.24	6.0E−58	0.14	3.0E−20	0.91	2.0E−37	0.8	0.92	0
	SD	Long	PC1	0.55	0	0.63	0	0.48	6.0E−253	0.21	8.0E−46	0.18	4.0E−32	0.92	2.0E−52	1.0	0.92	0
	SD	Long	PC2	0.57	0	0.68	0	0.48	3.0E−253	0.23	2.0E−54	0.17	4.0E−27	0.92	8.0E−56	1.1	0.94	0
	SD	Short	PC1	0.49	2.0E−264	0.62	0	0.43	3.0E−193	0.19	3.0E−37	0.12	2.0E−15	0.92	3.0E−42	0.9	0.89	0
	SD	Short	PC2	0.54	0	0.67	0	0.46	6.0E−222	0.22	3.0E−50	0.13	2.0E−18	0.91	2.0E−48	1.0	0.91	0

For relationships with ALSFRS-R and right-left agreement, p-values for Pearson’s correlation were computed using a Student’s t distribution for a transformation of the correlation (two-tailed test). Mann–Whitney U test was used for ALS versus control comparisons (two-sided test) and Cohen’s d was used to measure effect size. The Benjamini–Hochberg method was used to adjust for multiple comparisons and corrected p-values are reported.
Stat statistic, Dur duration, Dir direction, Dist distance, Vel velocity, Accel acceleration, SM submovement, M mean, SD standard deviation, AI activity intensity, PC principal component, ALSFRS-R ALS Functional Rating Scale-Revised, ICC intraclass correlation coefficient, r Pearson correlation coefficient, es effect size.

Back to article page

Table 1 Cross-sectional properties of ankle submovement models and features

Search

Quick links