Differences in running kinematics and stride-to-stride variability between self-paced treadmill and overground running on two surfaces

Treadmills are widely used in running research, yet it remains unclear how closely self-paced treadmill running reflects overground running. To date, treadmill-overground comparisons have been limited to fixed-speed conditions. This study compared self-paced treadmill running with overground running on hard and soft surfaces. Twenty-one physically active young adults performed running trials under three conditions: self-paced treadmill, overground soft surface, and overground hard surface. Mean values and stride-to-stride variability were analyzed for kinematic variables, and time-continuous sagittal kinematics were assessed using statistical parametric mapping. Compared with both overground conditions, treadmill running showed small but systematic differences in mean kinematics, including higher cadence (< 2 strides/min; p = 0.005), lower duty factor (< 3 %; p = 0.012), shorter stance time (< 22 ms; p = 0.002), smaller shank angle at initial contact (< 3°; p = 0.004), and reduced sagittal hip and knee ROM during stance (< 4°; hip: p < 0.001; knee: p = 0.001). Time-continuous analysis showed reduced hip and knee flexion (up to ∼ 10°), and greater ankle dorsiflexion (up to ∼ 9°) across large portions of the stride. Stride-to-stride variability was lower on the treadmill, with significant effects for cadence (p = 0.003), stance time (p = 0.019), shank angle, and sagittal hip, knee, and ankle ROM (all p < 0.001). These findings suggest that self-paced treadmill running does not fully replicate overground running and that treadmill-derived results, particularly for variability analyses, should be interpreted with caution in relation to overground running.