Fig. 8: Dimensionality of muscle coherence networks. | Communications Biology

Fig. 8: Dimensionality of muscle coherence networks.

From: Complexity of modular neuromuscular control increases and variability decreases during human locomotor development

Fig. 8

a Average (z-transformed, across subjects) percent of variance accounted for (VAF) the reconstruction of the inter-muscular coherence spectra between all muscle pairs in neonates (n = 13) and adults (n = 15) using different number of modules and different decomposition algorithms: principal component analysis (PCA, left) and non-negative matrix factorization (NNMF, right). Vertical error bars denote 95% confidence intervals across subjects. Data points for all individuals are included in the histograms. b Histograms of the number of modules that account for at least 80% of the VAF in neonates (upper row) and adults (lower row) using the two different algorithms (PCA left and NNMF right).

Back to article page