Evaluation of Spatiotemporal Patterns of the Spinal Muscle Coordination Output during Walking in the Exoskeleton

Abstract

Recent advances in the performance and evaluation of walking in exoskeletons use various assessments based on kinematic/kinetic measurements. While such variables provide general characteristics of gait performance, only limited conclusions can be made about the neural control strategies. Moreover, some kinematic or kinetic parameters are a consequence of the control implemented on the exoskeleton. Therefore, standard indicators based on kinematic variables have limitations and need to be complemented by performance measures of muscle coordination and control strategy. Knowledge about what happens at the spinal cord output level might also be critical for rehabilitation since an abnormal spatiotemporal integration of activity in specific spinal segments may result in a risk for abnormalities in gait recovery. Here we present the PEPATO software, which is a benchmarking solution to assess changes in the spinal locomotor output during walking in the exoskeleton with respect to reference data on normal walking. In particular, functional and structural changes at the spinal cord level can be mapped into muscle synergies and spinal maps of motoneuron activity. A user-friendly software interface guides the user through several data processing steps leading to a set of performance indicators as output. We present an example of the usage of this software for evaluating walking in an unloading exoskeleton that allows a person to step in simulated reduced (the Moon's) gravity. By analyzing the EMG activity from lower limb muscles, the algorithms detected several performance indicators demonstrating differential adaptation (shifts in the center of activity, prolonged activation) of specific muscle activation modules and spinal motor pools and increased coactivation of lumbar and sacral segments. The software is integrated at EUROBENCH facilities to benchmark the performance of walking in the exoskeleton from the point of view of changes in the spinal locomotor output.

Keywords: benchmarking; body unloading; exoskeletons; muscle coordination; spinal locomotor output; walking.

Figure 1

Overview of the spinal locomotor output evaluation system for walking in the exoskeleton. The acquired EMG signals and kinematic events serve as an input, and two main groups of performance indicators of the spinal muscle coordination output are generated as an output.

Figure 2

Pipeline and interfaces. (a) Pipeline of the PEPATO software for the EMG data analysis. (b) Examples of interfaces (screenshots) for the main elements of the data pre-processing chain: initial parameters selection (1), visualization of EMGs along with gait cycle timing and EMG spectra (2), exclusion of gait cycles with potential artefacts (3), identification of spectrum-related artefacts and their correction (4), warning for potential artefacts based on advanced artefact detection in individual cycles and their exclusion (5). (c) Cumulative VAF and FWHM, used for assessing the number of basic modules and the relative duration of basic pattern activity, accordingly. FWHM was calculated as the duration of the interval (in percent of gait cycle) in which EMG activity exceeded half of its maximum.

Figure 3

Examples of the EMG activity recordings during normal walking (a) and walking in the unloading exoskeleton (b) at 4 km/h. Experimental setup for recording of walking in the unloading exoskeleton is schematically shown on the left. Vertical lines correspond to the touchdown events in order to mark the individual gait cycles.

Figure 4

Assessment of performance indicators related to motor modules during walking in the unloading exoskeleton (UE) with respect to normal walking (NW). (a) Ensemble-averaged basic temporal patterns (±SD, left column) and corresponding muscle weights (right column, median and quartiles) of the two groups of subjects with four modules assumed for each group. Basic patterns were plotted in a “chronological” order (with respect to the timing of the main peak) vs. normalized gait cycle. (b) Biomechanical considerations: contributions of 4 basic patterns to the walking sub-tasks of body support, forward propulsion and leg swing during NW (modified from [40] with permission from Elsevier). Early stance (~15% of gait cycle), late stance (~45%), early swing (~70%), and late swing (~85%) are shown. Arrows departing from the center-of-mass denote the resultant module contributions to the horizontal and vertical ground reaction forces that accelerate the center-of-mass providing body support and forward propulsion. Net energy flow by each module to the trunk or leg is denoted by a + or - for energy increases or decreases, respectively. (c) Basic patterns similarity between UE and NW and similarity of corresponding muscle weights. (d) EMG data reconstruction (R²) using 4 modules, CoA of basic patterns, and FWHM. * Significant group differences.

Figure 5

Assessment of spatiotemporal maps of MN activity in the lumbosacral enlargement during walking in the unloading exoskeleton (UE) with respect to normal walking (NW). (a)—Examples of individual maps during normal upright walking (on the left) and walking in the unloading exoskeleton (on the right). Output pattern of each segment is shown in the top panels, while the same pattern is plotted in a color scale at the bottom. The pattern is reported in mV (since the EMG signal from each muscle was expressed in mV, see Materials and Methods). Motor output (averaged across several strides) is plotted as a function of gait cycle and spinal segment level (L2–S2). (b)—Performance indicators for the spinal maps (from left to right): similarities of sacral (S1 + S2) and upper lumbar (L3 + L4) segment activity during walking in UE with respect to NW, timing of maximum activity in the sacral and upper lumber segments, FWHM, and coactivation index. The values are plotted as median and quartiles. * Significant group differences are marked by asterisks.