Computer aided analysis of paraspinal electromyography

Hdl Handle:
http://hdl.handle.net/10149/301616
Title:
Computer aided analysis of paraspinal electromyography
Authors:
Coxon, A. (Andrew)
Advisors:
Longstaff, J. (Jim)
Citation:
Nnyanzi, L.A. (2012) The national child measurement programme: its value and impact. Unpublished PhD Thesis, Teesside University
Publisher:
Teesside University
Issue Date:
8-Feb-2013
URI:
http://hdl.handle.net/10149/301616
Abstract:
Back pain is responsible for British employees taking 5 million sick days per year. Low back pain (LBP) has a controversial aetiology, with 95% of cases caused by mechanical, non-pathological causes. Current medical treatment for mechanical LBP is an exercise regime designed to restore lumbar stability. Unfortunately this is often a painful process, and therefore difficult to complete. Electromyography (EMG) variables have been shown to be able to discriminate between subjects with and without mechanical LBP. If these variables could be shown to have discriminatory abilities before the actual onset of LBP they could be used to predict future episodes of LBP in currently otherwise asymptomatic individuals and allow the rehabilitation process to begin before the onset of symptoms. However a number of problems persist with EMG measurement. The test must be administered under closely controlled conditions in order to record clean signals, and interpretation of this data requires special tools and training. This thesis aims to make contributions in three main areas; AUTOMATED ANALYSIS Manual analysis of a large store of EMG raw data files is a time consuming process. If outcome variables that require manual interpretation are included this effect is magnified, with necessary questions being raised as to the accuracy and consistency levels that can be maintained. A successfully implemented automated system would reduce analysis time and improve confidence in the outcome variables recorded. Investigations will also be carried out into the addition of error detection and correction algorithms that could be performed during the analysis procedure. ECG CONTAMINATION REMOVAL Previous studies have identified ECG as a potential source of contamination of lumbar EMG signals. Compensation for this effect is non-trivial as the ECG frequencies overlap an area of interest in the EMG spectrum, and the ECG signal characteristics would change over a fatiguing EMG test. The Independent Component Analysis method will be used to attempt to extract and remove the ECG component of a recorded signal whilst preserving the underlying EMG data. If this is successful an analysis of the effect that removing ECG contamination has on EMG outcome variables will be presented. COLOUR MAP DIAGNOSTIC METHOD Colour maps are an excellent method of presenting a large amount of signal data to a researcher, and have been used to discriminate between LBP and non-LBP subjects. The usefulness of this diagnostic display too has been somewhat limited however by the difficulty in producing such maps. Investigations will be carried out into methods that will be able to quickly and accurately produce these colour maps to the same specification as earlier studies. Colour maps of subjects that did not report LBP at the time of testing, but who then did report LBP at their next presentation, will be examined to assess whether or not EMG colour maps can be used as a predictor for low back pain.
Type:
Thesis or dissertation
Language:
en
Keywords:
computer science; biomedical engineering; signal processing; diagnostic tools; electromyography; low back pain

Full metadata record

DC FieldValue Language
dc.contributor.advisorLongstaff, J. (Jim)en_GB
dc.contributor.authorCoxon, A. (Andrew)en_GB
dc.date.accessioned2013-09-16T14:26:51Z-
dc.date.available2013-09-16T14:26:51Z-
dc.date.issued2013-02-08-
dc.identifier.urihttp://hdl.handle.net/10149/301616-
dc.description.abstractBack pain is responsible for British employees taking 5 million sick days per year. Low back pain (LBP) has a controversial aetiology, with 95% of cases caused by mechanical, non-pathological causes. Current medical treatment for mechanical LBP is an exercise regime designed to restore lumbar stability. Unfortunately this is often a painful process, and therefore difficult to complete. Electromyography (EMG) variables have been shown to be able to discriminate between subjects with and without mechanical LBP. If these variables could be shown to have discriminatory abilities before the actual onset of LBP they could be used to predict future episodes of LBP in currently otherwise asymptomatic individuals and allow the rehabilitation process to begin before the onset of symptoms. However a number of problems persist with EMG measurement. The test must be administered under closely controlled conditions in order to record clean signals, and interpretation of this data requires special tools and training. This thesis aims to make contributions in three main areas; AUTOMATED ANALYSIS Manual analysis of a large store of EMG raw data files is a time consuming process. If outcome variables that require manual interpretation are included this effect is magnified, with necessary questions being raised as to the accuracy and consistency levels that can be maintained. A successfully implemented automated system would reduce analysis time and improve confidence in the outcome variables recorded. Investigations will also be carried out into the addition of error detection and correction algorithms that could be performed during the analysis procedure. ECG CONTAMINATION REMOVAL Previous studies have identified ECG as a potential source of contamination of lumbar EMG signals. Compensation for this effect is non-trivial as the ECG frequencies overlap an area of interest in the EMG spectrum, and the ECG signal characteristics would change over a fatiguing EMG test. The Independent Component Analysis method will be used to attempt to extract and remove the ECG component of a recorded signal whilst preserving the underlying EMG data. If this is successful an analysis of the effect that removing ECG contamination has on EMG outcome variables will be presented. COLOUR MAP DIAGNOSTIC METHOD Colour maps are an excellent method of presenting a large amount of signal data to a researcher, and have been used to discriminate between LBP and non-LBP subjects. The usefulness of this diagnostic display too has been somewhat limited however by the difficulty in producing such maps. Investigations will be carried out into methods that will be able to quickly and accurately produce these colour maps to the same specification as earlier studies. Colour maps of subjects that did not report LBP at the time of testing, but who then did report LBP at their next presentation, will be examined to assess whether or not EMG colour maps can be used as a predictor for low back pain.en_GB
dc.language.isoenen
dc.publisherTeesside Universityen
dc.subjectcomputer scienceen_GB
dc.subjectbiomedical engineeringen_GB
dc.subjectsignal processingen_GB
dc.subjectdiagnostic toolsen_GB
dc.subjectelectromyographyen_GB
dc.subjectlow back painen_GB
dc.titleComputer aided analysis of paraspinal electromyographyen_GB
dc.typeThesis or dissertationen
dc.publisher.departmentSchool of Computingen_GB
dc.type.qualificationnamePhDen
dc.type.qualificationlevelDoctoralen
or.citation.harvardNnyanzi, L.A. (2012) The national child measurement programme: its value and impact. Unpublished PhD Thesis, Teesside University-
cr.approval.ethicalSchool of Computing-
All Items in TeesRep are protected by copyright, with all rights reserved, unless otherwise indicated.