About

Graham Caldwell is an Associate Professor in the Department of Kinesiology at the University of Massachusetts Amherst. He holds a Ph.D. in Kinesiology with a specialization in Biomechanics from Simon Fraser University, obtained in 1987, as well as a Master of Science and a Bachelor of Science (Honours) in Kinesiology from the University of Waterloo, completed in 1980 and 1978 respectively. His research interests focus on the kinematics and kinetics of human movement, including the task-specific synergetic use of muscles and the computer modeling of muscle function. Caldwell utilizes musculoskeletal and forward dynamics models to study optimal movement patterns, contributing to the understanding of organismal biology and evolutionary biology within his academic discipline.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Mathematics
• Engineering
• Simulation
• Statistics
• Surgery
• Structural engineering
• Mathematical optimization
• Physical therapy
• Medicine
• Physical medicine and rehabilitation

Selected publications

• Lower Back Kinetic Demands During Induced Lower Limb Gait Asymmetries

  SSRN Electronic Journal · 2022-01-01

  articleOpen accessSenior author
  PublisherDOI

• Lower back kinetic demands during induced lower limb gait asymmetries

  Gait & Posture · 2022-09-05 · 6 citations

  articleSenior author
  PublisherDOI

• EMG optimization in OpenSim: A model for estimating lower back kinetics in gait

  Medical Engineering & Physics · 2022 · 29 citations

  Senior authorCorresponding
  • Computer Science
  • Physical medicine and rehabilitation
  • Computer Science

  Participant-specific musculoskeletal models are needed to accurately estimate lower back internal kinetic demands and injury risk. In this study we developed the framework for incorporating an electromyography optimization (EMGopt) approach within OpenSim (https://simtk.org/projects/emg_opt_tool) and evaluated lower back demands estimated from the model during gait. Kinematic, external kinetic, and EMG data were recorded from six participants as they performed walking and carrying tasks on a treadmill. For evaluation, predicted lumbar vertebral joint forces were compared to those from a generic static optimization approach (SOpt) and to previous studies. Further, model-estimated muscle activations were compared to recorded EMG, and model sensitivity to day-to-day EMG variability was evaluated. Results showed the vertebral joint forces from the model were qualitatively similar in pattern and magnitude to literature reports. Compared to SOpt, the EMGopt approach predicted larger joint loads (p<.01) with muscle activations better matching individual participant EMG patterns. L5/S1 vertebral joint forces from EMGopt were sensitive to the expected variability of recorded EMG, but the magnitude of these differences (±4%) did not impact between-task comparisons. Despite limitations inherent to such models, the proposed musculoskeletal model and EMGopt approach appears well-suited for evaluating internal lower back demands during gait tasks.

  PublisherDOI

• Muscle synergies are modified with improved task performance in skill learning

  Human Movement Science · 2022-03-22 · 21 citations

  articleSenior author
  PublisherDOI

• Are lower back demands reduced by improving gait symmetry in unilateral transtibial amputees?

  Clinical Biomechanics · 2022-04-26 · 5 citations

  articleCorresponding
  PublisherDOI

• A muscle control strategy to alter pedal force direction under multiple constraints: A simulation study

  Journal of Biomechanics · 2022-05-11 · 6 citations

  articleSenior author
  PublisherDOI

• A direct collocation framework for optimal control simulation of pedaling using OpenSim

  PLoS ONE · 2022 · 30 citations

  • Computer Science
  • Computer Science
  • Simulation

  The direct collocation (DC) method has shown low computational costs in solving optimization problems in human movements, but it has rarely been used for solving optimal control pedaling problems. Thus, the aim of this study was to develop a DC framework for optimal control simulation of human pedaling within the OpenSim modeling environment. A planar bicycle-rider model was developed in OpenSim. The DC method was formulated in MATLAB to solve an optimal control pedaling problem using a data tracking approach. Using the developed DC framework, the optimal control pedaling problem was successfully solved in 24 minutes to ten hours with different objective function weightings and number of nodes from two different initial conditions. The optimal solutions for equal objective function weightings were successful in terms of tracking, with the model simulated pedal angles and pedal forces within ±1 standard deviation of the experimental data. With these weightings, muscle tendon unit (MTU) excitation patterns generally matched with burst timings and shapes observed in the experimental EMG data. Tracking quality and MTU excitation patterns were changed little by selection of node density above 31, and the optimal solution quality was not affected by initial guess used. The proposed DC framework could easily be turned into a predictive simulation with other objective functions such as fastest pedaling rate. This flexible and computationally efficient framework should facilitate the use of optimal control methods to study the biomechanics, energetics, and control of human pedaling.

  PublisherOA PDFDOI

• Changes in Muscle Control After Learning to Direct Pedal Forces in One-Legged Pedaling

  Journal of Motor Learning and Development · 2021-05-06 · 3 citations

  articleSenior author

  The aim of this study was to describe how major leg muscle activities are altered after learning a novel one-legged pedaling task. Fifteen recreational cyclists practiced one-legged pedaling trials during which they were instructed to match their applied pedal force to a target direction perpendicular to the crank arm. Activity in 10 major leg muscles was measured with surface electromyography electrodes. Improved upstroke task performance was obtained by greater activity in the hip and ankle flexor muscles, counteracting the negative effects of gravity. Greater quadriceps activities explained improved targeting near top dead center. Reduced uniarticular knee and ankle extensor downstroke activities were necessary to prevent freewheeling. Greater hamstring and tibialis anterior activities improved targeting performance near the bottom of the pedal stroke. The activity patterns of the biarticular plantarflexors changed little, likely due to their contributions as knee flexors for smooth upstroke pedaling motion. These results add to our understanding of how the degrees of freedom at the muscle level are altered in a cooperative manner to overcome gravitational effects in order to achieve the learning goal of the motor task while satisfying multiple constraints—in this case, the production of smooth one-legged pedaling motion at the designated mechanical task demands.

  PublisherDOI

• Muscular activity patterns in 1-legged vs. 2-legged pedaling

  Journal of sport and health science/Journal of Sport and Health Science · 2020-01-20 · 8 citations

  articleOpen accessSenior author

  BACKGROUND: One-legged pedaling is of interest to elite cyclists and clinicians. However, muscular usage in 1-legged vs. 2-legged pedaling is not fully understood. Thus, the study was aimed to examine changes in leg muscle activation patterns between 2-legged and 1-legged pedaling. METHODS: Fifteen healthy young recreational cyclists performed both 1-legged and 2-legged pedaling trials at about 30 Watt per leg. Surface electromyography electrodes were placed on 10 major muscles of the left leg. Linear envelope electromyography data were integrated to quantify muscle activities for each crank cycle quadrant to evaluate muscle activation changes. RESULTS: Overall, the prescribed constant power requirements led to reduced downstroke crank torque and extension-related muscle activities (vastus lateralis, vastus medialis, and soleus) in 1-legged pedaling. Flexion-related muscle activities (biceps femoris long head, semitendinosus, lateral gastrocnemius, medial gastrocnemius, tensor fasciae latae, and tibialis anterior) in the upstroke phase increased to compensate for the absence of contralateral leg crank torque. During the upstroke, simultaneous increases were seen in the hamstrings and uni-articular knee extensors, and in the ankle plantarflexors and dorsiflexors. At the top of the crank cycle, greater hip flexor activity stabilized the pelvis. CONCLUSION: The observed changes in muscle activities are due to a variety of changes in mechanical aspects of the pedaling motion when pedaling with only 1 leg, including altered crank torque patterns without the contralateral leg, reduced pelvis stability, and increased knee and ankle stiffness during the upstroke.

  PublisherDOI

• Response to: Caution needed when interpreting muscle activity patterns during extremely low pedaling cadence

  Journal of sport and health science/Journal of Sport and Health Science · 2020-06-04

  letterOpen accessSenior author
  PublisherDOI

Recent grants

• NIH Grant R03AG026281

  NIH · $131k · 2009

Frequent coauthors

• Brian R. Umberger

  26 shared

• Richard E.A. van Emmerik

  University of Massachusetts Amherst

  23 shared

• Jacob J. Banks

  Harvard University

  22 shared

• Christopher J. Hasson

  Northeastern University

  21 shared

• Joseph Hamill

  University of Massachusetts Amherst

  20 shared

• Ross H. Miller

  19 shared

• Timothy R. Derrick

  Iowa State University

  12 shared

• Li Li

  Georgia Southern University

  10 shared

Education

• Ph.D., Kinesiology (Biomechanics)

  Simon Fraser University

  1987

• M.S., Kinesiology (Biomechanics)

  University of Waterloo

  1980

• Other, Kinesiology

  University of Waterloo

  1978