Functional Scaling of Musculoskeletal Models

Research output: Contribution to conference without publisher/journalConference abstract for conferenceResearchpeer-review


The validity of the predictions from musculoskeletal models depends largely on how well the morphology of the model matches that of the patient. To address this problem, we present a novel method to scale a cadaver-based musculoskeletal model to match both the segment lengths and joint parameters specific to the patient. This is accomplished using optimisation methods to determine patient-specific joint positions and orientations, which minimise the least-squares error between model markers and the recorded markers from a motion capture experiment.
Functional joint positions and joint axis orientations are then used to morph/scale a cadaver based musculoskeletal model using a set of radial basis functions (RBFs).
Using the functional joint axes to scale musculoskeletal models provides a better fit to the marker data, and allows for representation of patients with considerable difference in bone geometry, without the need for MR/CT scans. However, more validation activities are needed to better understand the effect of morphing musculoskeletal models based on functional joint parameters.
Translated title of the contributionFunktionel skalering af muskuloskeletale modeller
Original languageEnglish
Publication date2011
Number of pages2
Publication statusPublished - 2011
EventCongress of the International Society of Biomechanics, ISB - Brussels, Belgium
Duration: 3 Jul 20117 Jul 2011
Conference number: 23


ConferenceCongress of the International Society of Biomechanics, ISB

Bibliographical note

Paper available on USB memory stick.
Abstract in: Conference Book, Program & Abstracts, XXIIIrd Congress of the International Society of Biomechanics, ISB 2011, 3-7 July 2011, Brussels, Belgium, p. 96, No. IX.4,3. ISBN 9789090260198.

Fingerprint Dive into the research topics of 'Functional Scaling of Musculoskeletal Models'. Together they form a unique fingerprint.

Cite this