Comparison between kinematic and ground reaction force techniques for determining gait events during treadmill walking at different walking speeds.

This study involved a comparison of kinematics-based and ground reaction force (GRF)-based event detection methods. The objectives were (1) to assess the difference between the determination of gait events from GRF-based and kinematic/ultrasound-based techniques and (2) to determine the effects of gait speed on the agreement between the two techniques. At combined speeds, the average of the true difference was 2.6+/-2.8 ms for heel strike and -1.3+/-2.4 ms for toe-off identification. A positive value indicated that GRF-based identification occurred before kinematics-based identification. The average of the true difference was -3.9+/-3.5 ms for the duration of stance; thus, the duration of stance as determined by the kinematics-based technique was shorter than that determined by GFR-based detection. Strong correlations (range 0.948-0.974) were found at all (slow, moderate, and fast) gait speeds. Near unity of slope of the linear regression line (range 0.955-1.008) was identified for the duration of stance between the two methods. Our results suggest that the agreement between the two event detection methods depended on gait speed, but the differences were small. The data determined using kinematic/ultrasound-based and GRF-based methods were comparable in healthy participants. Copyright 2010 IPEM. Published by Elsevier Ltd. All rights reserved.