. 2018 Jun 3;2018:3795754. doi: 10.1155/2018/3795754

Table 3.

Study findings (cross-sectional designs).

Author, year	Study purpose	Primary findings	Findings related to module number, composition and control, gait, and rehabilitation outcomes
Allen et al. 2013 [25]	Determines biomechanical functions for modules used during poststroke walking using EMG data for simulations	Common merging patterns category A (modules 1 with 2) and category B (module 1 with 4) are correlated with common gait impairments seen in individuals poststroke	Category A: (1) Reduced propulsion during push off due to increased hamstring activity (2) Reduced limb swing (3) Reduced mediolateral stability during stance
Allen et al. 2013 [25]			Category B: (1) Reduced propulsion during push off due to premature plantarflexion (2) Reduced limb swing (3) Reduced body support during initial stance

Barroso et al. 2017 [34]	Determines if gait mechanics and modules are better predictors of walking performance than current clinical assessments (Fugl-Meyer)	Individuals poststroke used a range of 2–5 modules. There was a significant difference between mean VAF between the paretic and nonparetic limbs for individuals with 3, 4, or 5 modules (p < 0.018).	Stepwise multiple linear regression: Overground gait speed = 7.785–0.48 (time of peak nonparetic knee flexion) − 3.672 (nonparetic VAF with 4 modules) R² = 0.885, p = 0.001

Bowden et al. 2010 [26]	Determines the relationship of module use with gait mechanics and rehabilitation outcome measures	Module number was found to have higher correlations with functional outcomes and gait kinematics than the FM-LE or FMS	Gait kinematic correlations with module number Pp r = −0.389, p = 0.023 PSR r = −0.558, p = 0.001 PPS r = −0.398, p = 0.020
Bowden et al. 2010 [26]			Functional outcome correlations with module number Speed r = 0.451, p = 0.008 BBT r = 0.504, p = 0.003 DGI r = 0.545, p = 0.002

Clark et al. 2010 [14]	Determines if module use differences between healthy individuals and those poststroke are associated with walking performance	Number of modules used in the paretic limb during walking predicted performance	58% of participants required four modules 45% of participants required two modules 36% of participants required three modules
			Of 3 module participants, 8/19 demonstrated a merging of modules 1 and 2 (category A) and 7/19 demonstrated a merging of modules 1 and 4 (category B).
			Gait correlations with module use: SS speed p = 0.5, p = 0.0002 Speed modulation p = 0.47, p = 0.0008 Propulsive asymmetry p = −0.28, p = 0.04 Step length asymmetry p = −0.32, p = 0.02

Coscia et al. 2015 [27]	Evaluates relationships between gait asymmetries and changes in modules	Factor analysis resulted in a range of modules (3–5) to explain the gait cycle variance. The authors selected 3 modules as their maximum value which explained 75% of the variance.	In participants poststroke, module 1 explained a greater degree of variance than in healthy controls
Coscia et al. 2015 [27]			Changes in speed did not alter the number of modules, but did have a significant effect on weight coefficients for module 1 (p = 0.003) and module 2 (p = 0.027) between a stroke participant's paretic and nonparetic legs.

Gizzi et al. 2011 [29]	Determines if the number of modules is similar in individuals with subacute stroke (≤20 weeks) and if the inclusion of more muscles will change the number of modules	The number of modules was consistent with the previous findings in chronic stroke ranging from 2–4. The inclusion of upper extremity and trunk musculature did not change the number of modules.	Module number was equivalent using NNMF for the set of 16 muscles and set of 7 lower extremity only muscles.
Gizzi et al. 2011 [29]			Timing patterns for each module did not differ between limb-affected side (r = 0.74); unaffected side of patients, (r = 0.75); or healthy controls (r = 0.78, p = 0.05)

Kautz et al. 2011 [31]	Determines if there is a difference between TM and OG walking on module assignment	There was no difference between the numbers of modules assigned for individuals walking on the TM or OG despite differences in speed.	Module number explained greater than 90% variance for participants and controls for SS on TM (91.9% ± 4.1%, 93.5% ± 3.5%) OG (95.2% ± 2.8%, 97.5% ± 1.0%)
Kautz et al. 2011 [31]			Hemiparetic participants walked slower on the TM (TM, 0.38 versus OG, 0.58 m/s; p < 0.001), with increased cadence (TM, 84.9 versus OG, 77.6 steps/min, p = 0.04) and decreased stride length (TM, 0.52 versus OG, 0.85 m; p < 0.0001).

Routson et al. 2014 [33]	Determines if task variation results in module changes after a steady-state is reached	Varying SS speed walking conditions with maximum cadence, maximum step length, and maximum step height did not change the number of modules used.	In healthy controls, all tasks demonstrated 4 modules (p = 0.78).
Routson et al. 2014 [33]			Number of modules correlated to a reduced ability to change speed (p < 0.0001), cadence (p < 0.0001), step height (p < 0.0001), step length (p < 0.0001)

PC: principal component; FM-LE: Fugl-Meyer lower extremity; FMS: Fugl-Meyer synergy; BBT: Berg Balance Test; DGI: dynamic gait index; Pp: paretic propulsion; PSR: paretic step ratio; PPS: paretic preswing; SS: self-selected; FC: fastest comfortable; OG: over ground; TM: treadmill.