arXiv subjects

Dimitra Blana

Publications and source records attributed to Dimitra Blana.

3 recordsLinked to original sources

arXiv2026

Quaternion-Based Predictive Framework for Scapulohumeral Coordination

Scapulohumeral rhythm (SHR), the coordinated motion between the scapula and humerus during arm elevation, is frequently altered in rotator cuff pathologies, yet the mechanical principles underlying coordination redistribution remain difficult to explain from experimental data alone. This study presents a predictive optimal control framework for investigating scapulohumeral coordination, combining a quaternion-based shoulder model with EMG-informed muscle parameter calibration. The quaternion formulation eliminated kinematic singularities and associated non-physiological activation artifacts observed in the Euler-angle model, while maintaining comparable tracking accuracy. EMG-informed calibration reduced discrepancies between predicted and measured muscle excitations by up to 60% on independent validation tasks. In predictive simulations where only thoracohumeral elevation was prescribed, scapular and clavicular kinematics emerged from musculoskeletal mechanics and minimization of muscular effort, producing SHR values consistent with established experimental ranges. Simulated rotator cuff deficiency resulted in increased reliance on glenohumeral rotation. The proposed framework may serve as a tool for understanding impaired coordination patterns across a broad range of shoulder pathologies, with potential to inform personalized rehabilitation strategies and the design of assistive and prosthetic devices.

physics.med-ph↗

arXiv2025

Calculation of a force effect from muscle action to a quaternion-based musculoskeletal model

Euler angle representation in biomechanical analysis allows straightforward description of joints rotations. However, application of Euler angles could be limited due to singularity called gimbal lock. Quaternions offer an alternative way to describe rotations but they have been mostly avoided in biomechanics as they are complex and not inherently intuitive, specifically in dynamic models actuated by muscles. This study introduces a mathematical framework for describing muscle actions in dynamic quaternion-based musculoskeletal simulations. The proposed method estimates muscle torques in quaternion-based musculoskeletal model. Its application is shown on three-dimensional double-pendulum system actuated by muscle elements. Furthermore, transformation of muscle moment arms obtained from muscle paths based on Euler angles into quaternions description is presented. The proposed method is advantageous for dynamic modeling of musculoskeletal models with complex kinematics and large range of motion like the shoulder joint.

physics.med-ph↗

arXiv2024

Monkey Transfer Learning Can Improve Human Pose Estimation

In this study, we investigated whether transfer learning from macaque monkeys could improve human pose estimation. Current state-of-the-art pose estimation techniques, often employing deep neural networks, can match human annotation in non-clinical datasets. However, they underperform in novel situations, limiting their generalisability to clinical populations with pathological movement patterns. Clinical datasets are not widely available for AI training due to ethical challenges and a lack of data collection. We observe that data from other species may be able to bridge this gap by exposing the network to a broader range of motion cues. We found that utilising data from other species and undertaking transfer learning improved human pose estimation in terms of precision and recall compared to the benchmark, which was trained on humans only. Compared to the benchmark, fewer human training examples were needed for the transfer learning approach (1,000 vs 19,185). These results suggest that macaque pose estimation can improve human pose estimation in clinical situations. Future work should further explore the utility of pose estimation trained with monkey data in clinical populations.

cs.CV↗