Inconsistent Foot Contact Pattern Being Determined From Videos Of Lateral Or Medial Side During Running

Video-based methods are widely used to determine the pattern of initial foot-ground contact (FCP) at the beginning of the running stance phase, particularly during outdoor runs. The accuracy may be affected by camera position, i.e., on the medial or lateral side. PURPOSE: To investigate whether FCP is congruous when determined by videos from the medial or lateral side during running. METHODS: 20 males (20.8 ± 2.1 years old, 176.5 ± 4.6 cm, and 68.1 ± 5.8 kg) were recruited for 4 treadmill running trials in their comfort speeds (10.0 ± 1.5 km/h). For each trial, they ran in a random order for 3 min in a habitual or assigned FCP (rearfoot contact, RFC; midfoot contact MFC; forefoot contact, FFC). Ten markers were attached to each foot based on the foot model by Altman & Davis [1]. Marker trajectories were captured by a motion capture system with 12 high-speed infrared cameras, and videos were simultaneously recorded by two smartphones, respectively positioned at their lateral and medial sides. Marker data were processed, and the initial foot contact angle (FCA) was computed to determine FCP (Gold Standard). Videos were processed to obtain FCA and FCP. FCP was determined based on a widely used criteria that a MFC = −1.6° < FCA < 8.0°, a RFC = FCA > 8.0°, and a FFC = FCA < −1.6° [1]. Root-mean-square error (RMSE) was computed, and Pearson linear correlation tests were conducted to estimate accuracy. RESULTS: A total of 4800 strides were analyzed with 2704 (56.3%) RFC, 1224 (25.5%) MFC, and 872 (18.2%) FFC. FCA showed strong correlation with the overall RMSE of 2.91°-5.10° between video methods and gold standard. CONCLUSION: FCA had varied errors and FCP was inconsistent when measured by videos captured from lateral and medial sides. FCP determination was more reliable when camera was set at the lateral side during running. - Foot contact angle Running trial Overall(n = 4800) Forefoot contact Midfoot contact Rearfoot contact Habitual FSP Root Mean Square Error Lateral side VS. Gold standard 1.68° 2.12° 3.44° 3.83° 2.91° Medial side VS. Gold standard 3.38° 5.03° 5.73° 5.86° 5.10° Lateral side VS Medial side 3.81° 4.57° 4.62° 4.71° 4.44° Pearson Linear Correlation Lateral side VS. Gold standard r = 0.972 r = 0.934 r = 0.877 r = 0.870 r = 0.960 Medial side VS. Gold standard r = 0.775 r = 0.686 r = 0.700 r = 0.750 r = 0.887 Lateral side VS Medial side r = 0.782 r = 0.735 r = 0.799 r = 0.846 r = 0.922 Reference:[1] Altman, A.R. & Davis, I.S. Gait & Posture 2012; 35(2): 298-300.Supported by the China Scholarship Council (S.M.). Supported by: Supported by the China Scholarship Council (S.M.).