Muscle Synergies and Coherence Networks Reflect Different Modes of Coordination During Walking

Figures

• 

  FIGURE 1. Example of the determination of the modes of coordination based on kinematics of subject 1. (A) The average arm (blue) and ipsilateral leg (red) movement in the anterior-posterior (AP) direction as a function of the gait cycle in the 2:1 (1.0 km/h), transition (T, 2.5 km/h) and 1:1 (4.0 km/h) mode of coordination. (B) The spectral power. (C) The spectral overlap between the power spectra of the arm and leg is maximal for a 2:1 or 1:1 coupling between the arm and leg movement in the 2:1 and 1:1 mode of coordination, respectively. The transition contained peaks at both 2:1 and 1:1 coupling. (D) The relative phase between arm and leg. A generalized relative phase of zero slope implies that arm and leg move at a fixed frequency ratio (2:1, black and 1:1, gray).

• 

  FIGURE 2. Reconstruction quality of the muscle synergies. (A) Additional value of an extra synergy (Δλ) to the total contribution of the synergies to the Frobenius norm, (B) the contribution of every synergy (S1 to S5) to the Frobenius norm (λ). The order of synergies S1 to S5 is showed in Figure 3. Green, cyan and blue bar plots represent the 2:1, transition (T) and 1:1 mode of coordination, respectively.

• 

  FIGURE 3. Muscle synergies across modes of coordination. (A) The synergies’ temporal activation patterns as a function of the gait cycle derived from average muscle activity patterns for the different modes of coordination. Green, cyan and blue represent the 2:1, transition (T) and 1:1 mode of coordination. Error patches represent the standard error of the mean across subjects. (B) Synergies’ weightings across conditions and subjects in color scale. (C) Muscle synergy network plotted separately for each synergy on the body mesh (Makarov et al., 2015). A minimally-connected network was created for visualization. Node size represents the degree of the muscle and edge thickness represents weighted appearance of both muscles in the synergy.

• 

  FIGURE 4. Significant differences between synergies’ wave forms between modes of coordination. Green, cyan and blue represent 2:1, transition (T) and 1:1, respectively. Patches represent significant differences in time between the amplitude (gray) and the temporal patterns (red) of the synergies’ wave forms. α = 1/300.

• 

  FIGURE 5. The community structure of the multiplex (A) muscle synergy (B) and coherence networks based on the synergy and coherence spectra muscle weightings. Community structure is visualized by color-coded nodes and the average degree across layers of every muscle is displayed as node size on the body mesh (Makarov et al., 2015). The edge width is based on the average connectivity across layers between the muscles in either the minimally-connected synergy or coherence network.

• 

  FIGURE 6. Changes in connectivity between conditions and frequency components in the minimally-connected multiplex coherence network. (A) Frequency components 4–8, 8–22, and 22–60 Hz, obtained with non-negative matrix factorization. (B) Coherence networks in the 2:1, transition (T) and 1:1 mode of coordination (columns) and the frequency components (rows). Colors in the networks depict different modules and node size and edge width represent degree and connectivity strength between muscles, respectively. (C) Global efficiency, transitivity and average strength of the coherence networks per frequency component and condition. Error bars indicate standard errors of the mean and asterisks significant differences between conditions (α < 0.005).