Speed-dependent and mode-dependent modulations of spatiotemporal modules in human locomotion extracted via tensor decomposition

Abstract

How the central nervous system (CNS) controls many joints and muscles is a fundamental question in motor neuroscience and related research areas. An attractive hypothesis is the module hypothesis: the CNS controls groups of joints or muscles (i.e., spatial modules) by providing time-varying motor commands (i.e., temporal modules) to the spatial modules rather than controlling each joint or muscle separately. Another fundamental question is how the CNS generates numerous repertoires of movement patterns. one hypothesis is that the cnS modulates the spatial and/or temporal modules depending on the required tasks. It is thus essential to quantify the spatial modules, the temporal modules, and the task-dependent modulation of these modules. Although previous attempts at such quantification have been made, they considered modulation either only in spatial modules or only in temporal modules. These limitations may be attributable to the constraints inherent to conventional methods for quantifying the spatial and temporal modules. Here, we demonstrate the effectiveness of tensor decomposition in quantifying the spatial modules, the temporal modules, and the taskdependent modulation of these modules without such limitations. We further demonstrate that tensor decomposition offers a new perspective on the task-dependent modulation of spatiotemporal modules: in switching from walking to running, the CNS modulates the peak timing in the temporal modules while recruiting more proximal muscles in the corresponding spatial modules.How the central nervous system (CNS) controls the human body is a fundamental question. The human body possesses 360 joints and more than 650 muscles; the CNS controls the body while somehow resolving a significant number of degrees of freedom (DoFs), in fact, more than are necessary to achieve the desired motions 1 . For example, during walking and running in daily life, these large numbers of joints and muscles should be controlled in an orchestrated manner. The module hypothesis is an influential proposal regarding how such a tremendous number of DoFs can be managed 2-6 . According to this hypothesis, the CNS effectively reduces the number of DoFs to be managed by controlling groups of joints or muscles, referred to as spatial modules, rather than single joints or muscles separately. Accordingly, the time-varying motor commands sent to the spatial modules are referred to as temporal modules.How the brain constructs various repertoires of motions is another fundamental question. One possible solution is that the brain modulates the spatiotemporal modules depending on the task at hand. On the one hand, temporal modules, rather than spatial modules, may be modulated for specific tasks, such as walking, running, executing various types of gaits, or responding to unpredictable perturbations while walking 7,8 . On the other hand, task-dependent modulation may be applied to spatial modules rather than to temporal modules for certain tasks, such as responding to unpredictable perturbations to maintain bal...

Modulation of spatial and temporal modules in lower limb muscle activations during walking with simulated reduced gravity

“…Indeed, the spatial and temporal modules extracted in the two different types of techniques were related to each other (Figs. S1 and S2 ), as seen in a previous locomotion study 34 . Furthermore, the unweighted SP 4 and TE 4 modules in a gravity condition of 0.07 g were represented by the task-dependent variable close to 0 in the ST 3 module that was involved with SP 4 and TE 4 .…”

“…Thus far, we have discussed the gravity-dependent modulation of the modules at a roughly optimal walking speed corresponding to Fr ~ 0.25 in which the dynamic state of the body was geometrically similar among each gravity level because of the pendulum-like behaviour of the limbs 51 , 52 . However, not only gravitational load but also walking speed affect the generation of muscle activation patterns and their regulation based on modules 19 , 34 , 42 , 53 . As reported in previous studies, muscle activity at different walking speeds could be represented by spatial and temporal modules comparable with those at each walking speed 25 , 30 , 42 , in which task-dependent variables were systematically changed as a function of both gravity and treadmill speed in a related phase-dependent manner (Fig.…”

“…Matrix factorization of the EMG data enabled us to identify the differences in spatial and temporal modules in each gravity level, whereas how the spatial and temporal modules were modulated depending on the gravity levels could not be clearly demonstrated. Therefore, a tensor decomposition model (CANDECOMP/PARAFAC; CP) was fit to the EMG data to quantify the task-dependent variable in addition to the spatial and temporal (ST) modules 34 , 35 . For the tensor decomposition, the EMG data at Fr ~ 0.25 were arranged in a 3-dimensional array comprising 12 muscles × 200 time points × 5 gravity levels across all participants (i.e., 45).…”

“…Second, how pairs of spatial and temporal modules are modulated should be tested depending on the gravity levels. Tensor decomposition techniques make it possible to unravel this question by extracting task-dependent variables that vary in gravity, walking speed and participants in addition to fixed spatial and temporal modules 34 , 35 . The present results will provide a better understanding of neuromuscular control during walking under hypogravity conditions.…”

Effort-dependent effects on uniform and diverse muscle activity features in skilled pitching

“…1 ), tensor decomposition enables us to analyze three factors (i.e., the column [S], row [T], and number of slices in the third dimension [K]). Previous studies have shown the effectiveness of tensor decomposition and its variants in discussing task-dependent effects on spatiotemporal muscle modules 23 – 25 . We thus expect to discuss how spatiotemporal modules show effort-dependent effects by using tensor decomposition for each subject.…”