Accelerometers provide a method of collecting longitudinal running stride time (ST) and stride rate (SR) data; however, research is required to develop methods and analysis techniques which ensure the collection of valid data, interpreted correctly. The purpose of this thesis was to investigate and develop methods of collecting temporal gait data using accelerometry, in recreational distance runners. Furthermore, we wished to apply these methods to investigate SR, as a coaching orientated parameter, and ST variability, as a research orientated parameter.
Methods development was undertaken in which accelerometry self-attachment protocols and technical and processing issues which may affect ST calculation in accelerometry data were investigated. A novel method of ST calculation, utilising 2 Hz filtering, was developed which indicated good comparability to previous methods (ICCs > 0.95 and CV values < 1.5%). Following this, the methods developed were utilised to investigate SR in recreational runners during training runs, and both half and full marathon distances. Results indicated that performing with a comparable or decreased SR (compared to a similar training run) within a competitive distance running event may result in an optimal running style in relation to overall run outcome. The remaining sections of the thesis focused on ST as a research orientated parameter.
Long range correlations within running ST series have previously been linked with both injury and performance level, utilising the non-linear method Detrended Fluctuation Analysis (DFA). However, prior to implementing DFA on outdoor running time series, an investigation into how robust DFA is to outliers, which may occur due to perturbations or changes in terrain, was performed. Hereafter, DFA was applied to recreational runner’s half and full marathon ST series. No significant differences (p > 0.05), and small (ES = 0.35) and medium (ES = 0.46) effect sizes, were identified between recreational runner’s half marathon and full marathon DFA α values and SR, indicating that recreational runners adopt similar stride parameters when undertaking half and full marathon events, despite extended training periods. This information enhanced our knowledge of DFA during distance running, however due to extended processing times more efficient methods of utilising DFA were required. Therefore, a running analysis system was developed resulting in the output of reoccurring, real-time DFA α values during prolonged running. Preliminary investigation indicated output of DFA α values within 10 s of predetermined running time periods, enabling researchers to give real-time DFA feedback, during a prolonged run.
Overall this thesis represents a body of work encapsulating methods development, investigation into both coaching and research parameters of temporal gait and the development of a running gait analysis system. The research outlined has both sporting and clinical practical implications as the running gait methods developed can be utilised by future researchers and clinicians.