Modeling

MR Ebers, MC Rosenberg, JN Kutz, KM Steele (2023) “A machine learning approach to quantify individual gait responses to ankle exoskeletons”

Yusuke Maruo JournalArticle, News, Publications , ,

Journal Article in Journal of Biomechanics:

Physiological and biomechanical responses to mechanical assistance from wearable technology are highly variable, especially for clinical populations; tools to predict how users respond to different types of exoskeleton assistance may optimize the prescription process and uncover underlying mechanisms driving locomotor changes in the context of personalized wearable/assistive technology.

Aim: The purpose of this study was to determine if a discrepancy modeling framework could quantify individual-specific gait responses to ankle exoskeletons.

Method: We employ a machine learning technique — neural network based discrepancy modeling — on gait data from 12 non-disabled adults to capture within-participant differences in walking dynamics without vs. with a bilateral passive elastic ankle exoskeletons applying 5 N-m/deg of torque. We fit three models: Nominal gait (no exo), Exo, and Discrepancy. Then, post-fitting, we extend the Nominal by the Discrepancy Model (Augmented). We hypothesize that if Augmented (Nom+Discrep) can capture similar amount of variability as the Exo model, then it can be inferred that the discrepancy model accurately captures how a user will respond to an exoskeleton — without direct information about that user’s physiology or motor coordination.

Results:While joint kinematics during Exo gait were well predicted using the Nominal model (median 𝑅2 = 0.863 − 0.939), the Augmented model significantly increased variance accounted for (𝑝 < 0.042, median 𝑅2 = 0.928 − 0.963). For EMG, the Augmented model (median 𝑅2 = 0.665 −
0.788) accounted for significantly more variance than the Nominal model (median 𝑅2 = 0.516 − 0.664). Minimal kinematic variance was left unexplained by the Exo model (median 𝑅2 = 0.954 − 0.978), but only accounted for 72.4%–81.5% of the median variance in EMG during Exo gait across all individuals.

Interpretation:Discrepancy modeling successfully quantified individuals’ exoskeleton responses without requiring knowledge about physiological structure or motor control. However, additional measurement modalities and/or improved resolution are needed to characterize Exo gait, as the discrepancy may not comprehensively capture response due to unexplained variance in Exo gait.

MH Schwartz, KM Steele, AJ Ries, AG Georgiadis, BA MacWilliams (2022) “A model for understanding the causes and consequences of walking impairments”

Yusuke Maruo Energetics, JournalArticle, MotorControl, Publications ,

Journal Article in PLOS ONE:

Causal inference is inherently ambiguous since we cannot observe multiple realizations of the same person with different characteristics. Causal models must be evaluated through indirect means and reasoning.

Aim: The main objectives in conducting this study were to (1) propose a comprehensive model for quantifying the causes and consequences of walking impairments and (2) demonstrate the potential utility of the model for supporting clinical care and addressing basic scientific questions related to walking.

Method: This paper introduced a model consisting of 10 nodes and 23 primary causal paths and demonstrated the model’s utility using a large sample of gait data.

Results: The model was plausible, captured some well-known cause-effect relationships, provided new insights into others, and generated novel hypotheses requiring further testing through simulation or experiment.

Interpretation: This model is a proposal that is meant to be critically evaluated, validated or refuted, altered, and improved over time. Such improvements might include the introduction of new nodes, variables, and paths.

EC Kuska, N Mehrabi, MH Schwartz, KM Steele (2022) “Number of synergies impacts sensitivity of gait to weakness and contracture”

Katie JournalArticle, MotorControl, News, Projects, Publications , , , , , ,

Journal Article in Journal of Biomechanics

Muscle activity during gait can be described by a small set of synergies, weighted groups of muscles, that are theorized to reflect underlying neural control. For people with neurologic injuries, like cerebral palsy or stroke, even fewer synergies are required to explain muscle activity during gait. This reduction in synergies is thought to reflect altered control and is associated with impairment severity and treatment outcomes. Individuals with neurologic injuries also develop secondary musculoskeletal impairments, like weakness or contracture, that can impact gait. Yet, the combined impacts of altered control and musculoskeletal impairments on gait remains unclear.

A two-dimensional sagittal plane musculoskeletal model and synergy simulation framework tracked unimpaired gait kinematics. The model had nine degrees of freedom, including right and left leg hip, knee, and ankle flexion, actuated by eight muscles per leg. Fixed sets of synergies constrained control, forcing the direct collocation algorithm to solve for synergy activations. The objective function minimized deviations from unimpaired kinematics and the sum of muscle activations squared (neural demand). Weakness, simulated by a reduction in maximum isometric force, and contracture, simulated by a reduction in tendon slack length, were progressively increased for each muscle or muscle group until the simulation failed to replicate unimpaired gait. Kinematic deviations and convergence determined the success of the simulation. The primary outcomes were (1) musculoskeletal impairment thresholds, defined by the amount of weakness or contracture before failure, and (2) neural demand of each gait cycle.Aim: In this study, we use a two-dimensional musculoskeletal model constrained to synergy control to simulate unimpaired gait.

Methods: We vary the number of synergies, while simulating muscle weakness and contracture to examine how altered control impacts sensitivity to musculoskeletal impairment while tracking unimpaired gait.

Results: Results demonstrate that reducing the number of synergies increases sensitivity to weakness and contracture for specific muscle groups. For example, simulations using five-synergy control tolerated 40% and 51% more knee extensor weakness than those using four- or three-synergy control, respectively. Furthermore, when constrained to four- or three-synergy control, the model was increasingly sensitive to contracture and weakness of proximal muscles, such as the hamstring and hip flexors. Contrastingly, neither the amount of generalized nor plantarflexor weakness tolerated was affected by the number of synergies.

Interpretation: These findings highlight the interactions between altered control and musculoskeletal impairments, emphasizing the importance of measuring and incorporating both in future simulation and experimental studies.

MC Rosenberg, BS Banjanin, SA Burden, KM Steele (2020) “Predicting walking response to ankle exoskeleton using data driven models”

Elijah JournalArticle, News , , , , , ,

Journal Article in The Royal Society:

This work highlights the potential of data-driven models grounded in dynamical systems theory to predict complex individualized responses to ankle exoskeletons., without requiring explicit knowledge of the individual’s physiology or motor control

silhouette walking on left with purple lines and projections on right elipsoids and colored spheres

Aim: Evaluate the ability of three classes of subject-specific phase-varying (PV) models to predict kinematic and myoelectric responses to ankle exoskeletons during walking, without requiring prior knowledge of specific user characteristics.

Method: Data from 12 unimpaired adults walking with bilateral passive ankle exoskeletons were captured. PV, linear PV (LPV), and nonlinear PV (NPV) models leveraged Floquet theory to kinematics and muscle activity in response to three exoskeleton torque conditions.

Results: The LPV model’s predictions were more accurate than the PV model when predicting less than 12.5% of a stride in the future and explained 49–70% of the variance in hip, knee and ankle kinematic responses to torque. The LPV model also predicted kinematic responses with similar accuracy to the more-complex NPV model. Myoelectric responses were challenging to predict with all models, explaining at most 10% of the variance in responses.

Interpretation: This work highlights the potential of data-driven PV models to predict complex subject-specific responses to ankle exoskeletons and inform device design and control.

H Choi, TL Wren, KM Steele (2016) “Gastrocnemius operating length with ankle foot orthoses in cerebral palsy.” Prosthetics & Orthotics International

Dr. Steele JournalArticle, Orthoses , , ,
Example of gastrocnemius operating length from one subject with different AFOs.

Journal article in Prosthetics & Orthotics International:

How does the operating length of the gastrocnemius vary with different common AFOs in children with cerebral palsy?

Example of gastrocnemius operating length from one subject with different AFOs.Background: Many individuals with cerebral palsy wear ankle foot orthoses during daily life. Orthoses influence joint motion, but how they impact muscle remains unclear. In particular, the gastrocnemius is commonly stiff in cerebral palsy. Understanding whether orthoses stretch or shorten this muscle during daily life may inform orthosis design and rehabilitation.

Objectives: This study investigated the impact of different ankle foot orthoses on gastrocnemius operating length during walking in children with cerebral palsy.

Study design: Case series, within subject comparison of gastrocnemius operating length while walking barefoot and with two types of ankle foot orthoses.

Methods: We performed gait analyses for 11 children with cerebral palsy. Each child was fit with two types of orthoses: a dynamic ankle foot orthosis (Cascade dynamic ankle foot orthosis) and an adjustable dynamic response ankle foot orthosis (Ultraflex ankle foot orthosis). Musculoskeletal modeling was used to quantify gastrocnemius musculotendon operating length and velocity with each orthosis.

Results: Walking with ankle foot orthoses could stretch the gastrocnemius more than barefoot walking for some individuals; however, there was significant variability between participants and orthoses. At least one type of orthosis stretched the gastrocnemius during walking for 4/6 and 3/5 of the Gross Motor Functional Classification System Level I and III participants, respectively. AFOs also reduced peak gastrocnemius lengthening velocity compared to barefoot walking for some participants, with greater reductions among the Gross Motor Functional Classification System Level III participants. Changes in gastrocnemius operating length and lengthening velocity were related to changes in ankle and knee kinematics during gait.

Conclusion: Ankle foot orthoses impact gastrocnemius operating length during walking and, with proper design, may assist with stretching tight muscles in daily life.

Clinical relevance: Determining whether ankle foot orthoses stretch tight muscles can inform future orthotic design and potentially provide a platform for integrating therapy into daily life. However, stretching tight muscles must be balanced with other goals of orthoses such as improving gait and preventing bone deformities.