The effects of manipulating task constraints on running economy and multi-segment spine kinematics: Utilising a constraints-led approach to understand the impact of fatigue on running performance

Running performance is shaped by the interaction between physiological demand, biomechanical organisation and the ability to adapt movement under changing task constraints. Although running biomechanics has been widely investigated, the trunk has often been modelled as a single rigid segment, which limits understanding of how the thoracic, lumbar and pelvic regions contribute to running economy, fatigue and performance. This thesis investigated spinal and whole-body biomechanical and physiological adaptations during treadmill running using a constraints-led framework, validated multi-segment kinematic modelling, waveform-based analysis and single-subject interpretation.

The thesis is structured as a progressive programme of methodological and applied studies. First, a systematic review examined how three-dimensional spine, back and pelvis kinematics have been quantified during running using marker-based motion capture systems. The review identified substantial methodological variation across the literature, particularly in marker configuration, segment definition, coordinate-system construction, pelvic reporting, reliability evidence and outcome presentation. These findings highlighted the need for direct model evaluation before multi-segment spine models are used to interpret adaptation during running.

A model-comparison study then evaluated a four-segment skin-marker-set model and a three-segment cluster-based model during treadmill running in sub-elite male runners. Both models demonstrated excellent static criterion validity across the sagittal, frontal and transverse planes. However, during running, the models were not interchangeable. Agreement was strongest in the sagittal plane, partial in the frontal plane and weakest in the transverse plane, particularly in the lumbar region. These findings showed that model choice can alter the magnitude, timing and regional distribution of measured spinal motion. A subsequent reliability study demonstrated that the selected skin-marker-set model produced acceptable between-session repeatability for most segment-plane combinations, supporting its use in repeated-measures running research, although lumbar sagittal range of motion and upper thoracic axial rotation required cautious interpretation.

The applied studies then used this model to examine how sub-elite male runners adapted to acute localised lower-limb fatigue and increased running intensity. Across these task constraints, runners generally preserved the global structure of gait, but reorganised local spinal behaviour through selective changes in segmental motion, waveform timing, segmental dominancy and coupling-angle variability. Acute fatigue did not produce a uniform breakdown in running mechanics. Instead, adaptation was region-, plane- and participant-specific, with the greatest sensitivity often observed in middle spinal couplings and around the lumbopelvic interface. Physiologically, fatigue and increased intensity altered oxygen cost and ventilatory demand, but these changes did not map onto one consistent biomechanical response across all runners.

Increased running intensity produced clearer temporal-spatial and transverse-plane demands, yet individual responses remained central to interpretation. Some runners showed relatively stable coordination patterns, whereas others demonstrated altered coupling behaviour and incomplete recovery when returning to the initial lower-intensity running condition. This suggests that adaptation to running constraints is not simply a linear shift from one stable pattern to another, but a dynamic process in which runners preserve task performance by reorganising local coordination strategies.

Overall, this thesis demonstrates that spinal behaviour during running is adaptive, relational and individual. The findings show that meaningful adaptation may be missed when the trunk is treated as a single rigid segment or when interpretation relies only on group-level discrete variables. By combining validated multi-segment modelling, waveform analysis, vector coding and single-subject interpretation, this thesis provides methodological evidence for reliability-informed spine analysis and empirical evidence that trained runners maintain global running performance through local reorganisation of spinal coordination. These findings contribute to a more nuanced understanding of running economy, fatigue and biomechanical adaptation under changing task constraints.