Gait impairments are commonly seen after a stroke, affecting more than 70% of stroke survivors who usually exhibit hemiparetic patterns of weakness1. The ability to independently ambulate a distance of 10 m is indicative of patients’ lower limb function and overall motor performance in post-stroke recovery2, and walking speed is often used to evaluate gait performance post stroke in clinical settings3. Apart from walking speed, clinimetric and functional assessments have been applied clinically to evaluate gait performance, lower extremity joint strength, and muscle force. These are important predictors for post-stroke ambulatory function3–6.
To better understand the underlying nature of the post-stroke hemiplegic gait, biomechanical measurements have also been extensively studied, in particular, spatiotemporal parameters (e.g., walking speed, step length, step width)7,8, joint range of motion (ROM) or maximum joint angles in various gait phases7,9,10, and joint angles at specific gait events of the affected (paretic) body side7. While the existing literature focuses on the affected limb, which displays abnormal kinematic variables, specific descriptions of kinematic or kinetic of the contralateral (less affected) limb are sparse. Given the possible biomechanical adaptions of the less affected limb, precise measurements may provide insights towards understanding the gait patterns of the contralateral limb11.
Recent research has highlighted the significance of correctly measuring the nature of impairment and disability in heterogenous stroke populations, with the aims to prescribe individualized and effective treatments12. Recent technologies such as motion capture as assessment tools can systematically examine gait deviation and track rehabilitation outcomes more accurately. While the abovementioned studies have examined the discrete (zero- dimensional, 0D) variables in stroke survivors using instrumented gait analysis systems or (3D) motion capture devices7–10, information about the time-history of these biomechanical variables in a full gait cycle is unclear.
Statistical Parametric Mapping (SPM), as a statistical analysis tool, is able to detect differences in one-dimensional (1D) biomechanics data (e.g., time-varying waveforms for forces, joint angles, joint moments, electromyography amplitudes) between two or more conditions/groups13,14. This method has been applied in gait analysis for able-bodied and athletic populations15,16. For stroke survivors, it has not been widely utilized. A recent study examined the gait variables using SPM in hemiplegic gait, and observed greater thorax flexion/extension angle during stance phase and greater thorax internal/external rotation angle during the terminal stance phase in the stroke group than the control group7.
When applying SPM in gait analysis, many previous studies treated the gait cycle as an entire phase (0 to 100%). While some reported significant differences in certain periods of the stance or swing phase, e.g., ‘during the pre-swing and initial swing phases (55.2–66.5%)’17 and ‘terminal-stance phase (31–50%)’18, it is difficult to accurately identify the stance and swing phases in the group waveforms. These studies may have used the cut-off value of 60% to split an entire gait cycle (stance phase: 0 to 60%, swing phase: 60–100%)19. However, it can be more problematic for stroke survivors, who usually a display prolonged stance phase and a higher ratio of stance duration to swing duration in an entire gait cycle20. Hence, dividing an entire gait cycle is warranted in SPM analysis7, which can provide more detailed information regarding abnormal gait patterns. This present study, therefore, aimed to apply SPM to compare the biomechanical variables of both the affected and less affected limbs in the stance phase and swing phase of a gait cycle during a 10-m walking task between the stroke and control groups.