Table 2.

Smartphone sensor features from each active and passive monitoring activity and corresponding MDS‐UPDRS item(s) score, associated test‐retest reliability, ability to detect control vs. PD differences, relationship with clinical symptom severity and ability to detect signs in PD participants clinically scored as having no signs

Activity	Sensor Feature and Unit	MDS‐UPDRS Item(s)	Test‐Retest Reliability of Sensor Feature	Controls vs. All PD		PD Participants: Relationship Between Active Test Feature and MDS‐UPDRS Item(s) Score		Controls vs. PD Subgroup Rated as Having No Symptoms (MDS‐UPDRS score = “0”) at Respective Site Visit
Active tests			ICC	t Value	P Value	t Value	P Value	t Value	P Value
Sustained phonation	MFCC2	2.1	0.82	−4.48	<0.001	–0.36	n.s	−3.29	<0.001
Rest tremor	Skewness	3.18	0.90	7.6	<0.001	2.17	0.033	5.5	<0.001
Postural tremor	Power (m2/s3)	2.10	0.97	4.42	<0.001	2.61	0.011	0.83	n.s.
Finger tapping	Tap variability	2.5	0.64	8.46	<0.001	2.18	0.028	6.95	<0.001
Balance	Mean velocity (m/s)	3.13	0.80	8.29	<0.001	2.38	0.027	5.16	0.005
Gait	Turn speed (degree/s)	PIGD	0.88	−9.31	<0.001	−2.45	0.022	−8.25	<0.001
Passive monitoring			ICC	Mann‐Whitney U	P Value	t Value	P Value	Mann‐Whitney U	P Value
Walking	Turn speed (degree/s)	Average PIGD	n/a	1,392	<0.001	−3.01	0.005	229	0.018
Sit‐to‐stand transitions	Sit‐to‐stand transitions (#/hour)	Average PIGD	n/a	1,087	0.005	−1.48	n.s.	250	n.s.
Walking vs. not walking	Activity ratio (%)	Average PIGD	n/a	1,392	<0.001	−2.02	0.055	321	0.003

Values for the active tests are t values from linear mixed‐effects models that quantify the relationship between given MDS‐UPDRS item(s) and the selected features and take repeated tests over time of the same subjects into account. Values for the passive monitoring activities are U values from the Mann‐Whitney U test.

MFCC2, mel‐frequency cepstral coefficient 2; n/a, not applicable; n.s., not significant; STS, sit‐to‐stand/stand‐to‐sit transitions.