Feasibility of EMG-Based Control of Shoulder Muscle FNS Via Artificial Neural Network

Abstract

We investigated the potential use of EMG recordings from voluntary shoulder muscles in individuals with C5 spinal cord injury to automatically control the stimulation to paralyzed shoulder muscles in a task-appropriate manner. A musculoskeletal model of the human shoulder and elbow was modified to have maximum muscle forces appropriate for C5 spinal cord injury, including completely and partially paralyzed muscles. Inverse model simulations generated muscle activation levels that were used to train an artificial neural network (ANN) to automatically generate appropriate stimulation patterns for the "paralyzed" muscles based on "voluntary" muscle activations. We found that substantial additional shoulder strength could be provided by assuming that just two paralyzed muscles (pectoralis major and latissimus dorsi) were stimulated. Further, the needed activations of these "stimulated" muscles could be predicted with reasonable accuracy using the activation levels just two "voluntary" muscles (trapezius and rhomboids) as ANN inputs.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 25, 2001
Accession Number
ADA411436

Entities

People

  • A. M. Acosta
  • F. Van Der Helm
  • P. P. Parikh
  • R. F. Kirsch

Organizations

  • Case Western Reserve University

Tags

Communities of Interest

  • Air Platforms

DTIC Thesaurus Topics

  • Arm Bones
  • Biomechanical Phenomena
  • Biomedical Engineering
  • Body Weight
  • Elevation
  • Engineering
  • Errors
  • Military Research
  • Neural Networks
  • Paralysis
  • Shoulder
  • Simulations
  • Spinal Cord
  • Spinal Injuries
  • Three Dimensional
  • Upper Extremity

Readers

  • Computational Modeling and Simulation
  • Exercise and Sports Science.
  • Neuroscience

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks