Loading...

Stroke Neuromusculoskeletal Modeling in Order to Understand Rehabilitative Interventions

Hajihosseinali, Majid | 2023

67 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 55793 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Behzadipour, Saeed; Farahmand, Farzam
  7. Abstract:
  8. This dissertation focuses on modeling human hand movements post-stroke to understand rehabilitative interventions. The model consists of a multi-layer neural network as the motor cortex in the brain and a two-degree freedom biomechanical model with six muscles for the simulation of hand movement in the horizontal plane. A passive robotic device has been designed and manufactured to record planar hand movements. Movements during reaching tasks have been recorded in 24 chronic stroke patients, as well as 18 healthy subjects, using the robotic device. Statistical analysis of the data showed that movements in a particular direction (NW-SE) are more sensitive to stroke, improve after rehabilitation more than other directions, and are more correlated with clinical criteria with respect to the other directions. This finding proposes a clinical hint: Movements in the NW-SE direction are the best direction for the evaluation of stroke patients. Movement data has been used for training the model in three scenarios: healthy subjects, stroke patients before rehabilitative interventions, and stroke patients after interventions. Simulation results showed that the model can predict accurately the movement path of healthy subjects. Furthermore, neuron activation in the healthy scenario displays a preferred direction that almost corresponds to the NW-SE direction. This can explain why the NW-SE direction is more sensitive: Movement in the NW-SE direction is more difficult, because of more neuron and muscle activations, therefore this direction is more sensitive to stroke and rehabilitation exercises. Simulation results in a patient before intervention showed that neuron preferred direction has been turned clockwise with respect to healthy subjects. Although the model of this patient shows counterclockwise rotation after rehabilitation, it differs from healthy subjects. Therefore, neuron preferred direction can be proposed as a quantitative criterion for the level of disability in stroke patients
  9. Keywords:
  10. Musculoskeletal Modeling ; Neural Network ; Rehabilitation ; Stroke ; Rehabilitation Exercises ; Motor Cortex Modeling

 Digital Object List

 Bookmark

No TOC