How many strides are required for a reliable estimation of temporal gait parameters? Implementation of a new algorithm on the phase coordination index

Background: The Phase coordination index (PCI), a temporal gait measure that quantifies consistency and accuracy in generating the anti-phased left-right stepping pattern, assesses bilateral coordination of gait in various cohorts (e.g., Parkinson's disease, post stroke). As PCI is based on mean values calculated across a series of gait cycles, individuals are required to perform lengthy walking trials, prolonging gait assessments which cause discomfort to some of them. This study introduces an algorithm to identify the required number of strides to obtain a reliable, characteristic PCI value. Methods: Simulated data sets, as well as physiological data (obtained from healthy elderly and young persons, from over ground and treadmill trials) were used in this research. A series of N-1 PCI values was calculated for i = 2,3,4…N gait cycles for each participant. There is a value i = k, representing certain number of cycles, for which no significant change in PCI occurs as additional cycles are added, termed point of stabilization (POS). The algorithm presented here uses a 2-stage iterative process to determine POS. Stage 1 searches for the gross location of the interval of PCI values containing the POS. In stage 2, the algorithm performs a high-resolution recursive, iterative process within this interval to find the exact point. The criterion for defining stability within a window of PCI values is a coefficient of variation (CV) of ≤ 5%. Results: Our recursive, iterative algorithm indicates that ~23 strides on average should be captured to attain a characteristic PCI. Conclusions: Gait trials with at least 23 strides on average should suffice to obtain a reliable estimation of PCI in healthy young adults. While this methodology may be considered generic, future studies should obtain POS values based on additional cohorts (e.g., disabled participants, fixed walking speeds).