Computational modelling techniques to determine patellofemoral joint reaction forces

De Jongh, Bruce (2016-03)

Thesis (MEng)--Stellenbosch University, 2016.

Thesis

ENGLISH ABSTRACT: A computational modelling pipeline is introduced to determine knee muscle forces, knee kinematics, patellofemoral contact force and patella kinematics during stance phase of over ground level walking. By applying a computational method the underlying causative factors of anterior knee pain can be investigated and can be used to assist physicians in developing a treatment plan. Eleven subjects with anterior knee pain were tested pre- and post-physiotherapeutic intervention at the Central Analytical Facility Motion Analysis Clinic to record the kinematic and kinetic data of the body. One force plate is available to measure ground reaction force (GRF) data; therefore the GRF on the foot not striking the force plate is estimated using two methods, which are compared. Within the method used in this project, an average GRF and centre of pressure (COP) is computed throughout stance phase using force plate data and is then superimposed onto the foot not striking the force plate. The knee muscle forces and joint kinematics are then computed in OpenSim, whereby a generic musculoskeletal model is scaled according to the subject’s anthropometric data, where after the subjects kinematic and kinetic data, recorded in the motion laboratory, is applied to the model. The muscle forces compare well to literature in terms of magnitude and their linear relationship, however, the division of the forces between the muscles of the quadriceps femoris does not agree with literature. This is as a result of the inaccurate hip adduction and rotation angles and the simplified knee joint. Furthermore, errors arise in the modelling pipeline as a result of the superimposition method. The difference in magnitude of the GRF measured on the different feet and between sessions, results in knee moments and muscle forces that differ in magnitude. The computed muscle forces and knee kinematics are applied to a patient specific musculoskeletal model generated by segmenting an MRI scan of the affected knee. The patella kinematics and patellofemoral contact force were investigated during stance phase. The patella was situated more laterally when the knee was relaxed and when the knee begins to flex it moves medially until trochlear engagment, where after its lateral displacement remains constant. The initial medial displacement is caused by the medial patellofemoral ligament (MPFL) that applies considerable tension to the patella during trochlear engagment. The patellofemoral conact force agrees with values published in literature and computes repeatable results within sessions. The small difference between sessions and the contact force in literature is due to the rectus femoris force applied to the patella. Knee and hip adduction and rotation effect the patella kinematics and therefore have an effect on the mediolateral contact force. Due to limitations such as having one force plate and renovations being done on the Motion Analysis Clinic during this project, accurate results were not compared pre- and post-intervention. However, the modelling pipeline is validated as it produced repeatable results in the majority of the steps of the pipeline. If results were not repeatable, the cause could be attributed to the assumptions made in this project.

AFRIKAANSE OPSOMMING: Die gebruik van ʼn numeriese modellerings tegniek word beskryf waarvolgens knie spier kragte, knie kinematika, patella-femorale kontak krag en knieskyf kinematika tydens die standsfase in normale loop beraam kan word. Toepassing van die tegniek bemoontlik die ondersoek na onderliggende faktore wat kan lei tot anterior kniepyn met die doel om dokters by te staan in die ontwikkeling van nuwe behandelingsplanne. Elf deelnemers met anterior kniepyn was getoets voor en na fisioterapeutiese behandeling by die Sentrum vir Analitiese Fasiliteite se bewegingsopname kliniek om kinematika en kinetika opnames van die lyf te neem. Die grond reaksiekrag data van die voet wat nie die lassel tref nie is benader m.b.v. twee metodes wat teen mekaar opgeweeg word. Die metode wat in die modelleringspyplyn gebruik was behels die beraming van die gemiddelde grond reaksiekrag en drukmiddelpunt soos bereken vanaf die lassel data van die teenoorstaande voet stand fase. Die beraamde parameters word dan op die voet wat nie die lassel tref nie aangewend. Deur om die eksperimentele data vanaf die bewegingskliniek te gebruik, is ʼn generiese OpenSim model geskaleer volgens die deelnemers se antropometries mates en is gebruik om die knie spierkragte en gewrig kinematika te bereken. Die berekende spierkragte vergelyk goed met literatuur i.t.v. omvang sowel as tendens. Die onderverdeling van die kragte tussen die spiere stem egter nie ooreen met literatuur nie. Die krag verdeling tussen die kwadrisep spier is verkeerd en dit kan toegeskryf word aan die onakkurate heup adduksie en rotasie hoeke asook die vereenvoudigde een graad van vryheid knie. Foute sluip ook in a.g.v. die superponerings metode. Die verskil in omvang van die gemete en grond reaksiekrag tussen voete en sessies lei tot knie momente en spierkragte wat verskil in omvang. Die berekende spierkragte en knie kinematika is ingespan in ʼn pasiënt spesifieke spier-skelet model soos beraam vanaf gesegmenteerde MRI beelde van die geaffekteerde knie. Die knieskyf kinematika en patella-femorale kontak krag is ondersoek vir die standfase. Die knieskyf was in ʼn laterale posisie by volle ekstensie waarna dit mediaal begin skuif aan die begin van fleksie tot en met trochlea interaksie. Die laterale verplasing bly daarna konstant. Die inisiële mediale verplasing kan toegeskryf word aan die mediale patellofemorale ligament wat noemenswaardige tensie op die knieskyf uitoefen voor trochlea interaksie. Die patello-femorale kontak krag is soorgelyk aan waardes soos vervat in die literatuur en berekende waardes is herhaalbaar tussen sessies. Die klein verskille tussen sessies en die kontak krag in die literatuur is a.g.v. die krag wat deur die rectus femoris spier aangewend word. Knie en heup adduksie en rotasie beïnvloed knieskyf kinematika en daarom het dit ook ʼn effek op die mediaal-laterale kragte. v A.g.v. beperkings, soos bv. een lassel en die feit dat verbeteringe aan die bewegingskliniek aangebring is tydens die uitvoering van die projek, kon akkurate vergelykings tussen die voor en nabehandeling nie geskied nie. Die herhaalbaarheid van die metode kon egter bewys en bevestig word vir meeste stappe in die modelleringspyplyn en waar daar verskille was, kon dit toegeskryf word aan die aannames wat gemaak is.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/98770
This item appears in the following collections: