Table 1.

Papers about early diagnosis.

References	Tech.	Sensor place	Rec. freq	Experimental design	Subjects	Feature extracted	Analysis/classifiers	Classifier performance or findings
Chen et al., 2014	ACC, GYRO	Trunk	Not reported	Upright standing position, arms crossed on the chest, looking ahead (each 30 s): (i) eyes open (EO), (ii) feet together eyes closed (EC), (iii) feet together EO dual task (EODT), (iv) EC dual task (ECDT), (v) tandem standing EO (TEO)	24 PwPD -SBD, 23 HC	RMS ACC and Jerkiness of sway (Jerk) in mediolateral (ML) and anteroposterior (AP) directions	ANOVA, t-Test, Pearson Chi Square test	Differences in Jerk between PD/HC for EODT (p = 0.030), ECDT (p = 0.015) and TEO (p = 0.023)
Mancini et al., 2011	ACC, GYRO	Trunk	100 Hz	Posture: (i) EO gaze straight ahead at an art poster 6 m ahead; (ii) EC, upright standing position; (iii) EC, cognitive task (ECT)	13 PwPD de novo, 12 HC	Jerk, RMS of displacements, mean velocity, frequency (freq) below which is 95% of the power of the COP displacement power spectra (F95%), freq dispersion (FD)	Linear mixed model; ROC	AUC: 0.90 for F95%, 0.87 for FD, 0.82 for RMS, 0.93 for Jerk (EO) for untreated PD/HC classification. No significant correlation with UPDRS III
Sant'Anna et al., 2011	GYRO	Wrists, shanks	200 Hz	Walking 30 m hallway, preferred speed (2 min)	11 early/mild PwPD, 15 HC	Symmetry indexes	t-Test, ROC; ICC	0.872 AUC for PD/HC classification; ICC = 0.949
Perumal and Sankar, 2016	Force sensors	Below toe, toe	100 Hz	Gait and Tremor (60 s)	Gait: 93 PwPD (HY 2/3), 73 HC; Tremor: 16 PwPD	Gait: Step length, stride time, stance time, swing time, heel, below toe and toe forces and normalized heel, below toe and toe forces. Tremor: amplitude, power distribution, FD, median freq	ANOVA; LDA 5-fold cross validation; ROC	Gait: mean acc.: 91.58%, ROC 0.72 sens., 0.81 spec. and AUC 96% for PD/HC classification. Tremor: features able to distinguish PD tremor from atypical PD tremor