Use this URL to cite or link to this record in EThOS:
Title: Ultrasonography for the prediction of musculoskeletal function
Author: Miguez, Diego
ISNI:       0000 0004 7226 6940
Awarding Body: Manchester Metropolitan University
Current Institution: Manchester Metropolitan University
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
Ultrasound (US) imaging is a well-recognised technique for studying in vivo characteristics of a range of biological tissues due to its portability, low cost and ease of use; with recent technological advances that increased the range of choices regarding acquisition and analysis of ultrasound data available for studying dynamic behaviour of different tissues. This thesis focuses on the development and validation of methods to exploit the capabilities of ultrasound technology to investigate dynamic properties of skeletal muscles in vivo exclusively using ultrasound data. The overarching aim was to evaluate the influence of US data properties and the potential of inference algorithms for prediction of net ankle joint torques. A fully synchronised experimental setup was designed and implemented enabling collection of US, Electromyography (EMG) and dynamometer data from the Gastrocnemius medialis muscle and ankle joint of healthy adult volunteers. Participants performed three increasing complexity muscle movement tasks: passive joint rotations, isometric and isotonic contractions. Two frame rates (32 and 1000 fps) and two data precisions (8 and 16-bits) were obtained enabling analysis of the impact of US data temporal resolution and precision on joint torque predictions. Predictions of net joint torque were calculated using five data inference algorithms ranging from simple regression through to Artificial Neural Networks. Results indicated that accurate predictions of net joint torque can be obtained from the analysis of ultrasound data of one muscle. Significantly improved predictions were observed using the faster frame rate during active tasks, with 16-bit data precision and ANN further improving results in isotonic movements. The speed of muscle activation and complexity of fluctuations of the resulting net joint torques were key factors underpinning the prediction errors recorded. The properties of collected US data in combination with the movement tasks to be assessed should therefore be a key consideration in the development of experimental protocols for in vivo assessment of skeletal muscles.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available