Table 2.

Methodological characteristics of the included studies.

	Participants and Time of Inclusion	Task	Timepoints	Methods	Dependent Variables
Caty et al. (2009) [30]	n = 10, >6-month post-stroke	2 min treadmill walk	3 measures: test repeated 1 day and 1 month after baseline	Optical motion capture and oxygen consumption measures	Lower limb kinematic variables
Lofrumento et al. (2021) [31]	n = 6, <1-month post-stroke;	3 min treadmill walk and 6 steps overground	Tests before and after 4 weeks of conventional therapy	Optical motion capture with UCM analysis method a	Variance of UCM and the orthogonal subspace of the ankle joint trajectories
	n = 4, >6-month post-stroke
Papi et al. (2015) [32]	n = 1, 2 months post-stroke;	6 repeats of 6 meters of overground walking	3 measures: test repeated 3 and 6 months after baseline	Optical motion capture with UCM analysis method a	Variance of UCM and the orthogonal subspace of lower limb sagittal joint kinematics
	n = 6, healthy controls
Shin et al. (2020) [33]	n = 9, <1-month post-stroke;	Measures during conventional therapy	5 to 12 sessions over the course of rehabilitation	Inertial motion capture to extract amount of activity information	Amount of motion (total amount of joint displacements measured from inertial motion capture)
	Subgroup of n = 6 was longitudinally monitored
Roby-Brami et al. (2003) [34]	n = 6, single measure, 48–162 days post-stroke;	Seated reaching movements	1 to 3 measures at monthly timepoints during regular therapy	Electromagnetic motion tracking	Peak velocity of the hand and movement duration, amount of acromion displacement, and joint angular variations
	n = 9, repeated measures, 24–89 days post stroke;
	n = 7 healthy controls

^a UCM: uncontrolled manifold [35,36].