Investigation of a Double Calibration Technique to Reduce Soft Tissue Artifact Error in High Flexion

Abstract

High knee flexion tasks (knee flexion angle > 120 degrees) are performed frequently in both daily living activities (gardening, religious practice, exercise, etc.) and occupational settings (childcare, roofing, construction, floor laying, etc.). These tasks are associated with an increased risk of knee osteoarthritis development, which can alter movement patterns. These types of movement differences can be captured when analyzing high flexion postures using optical motion capture; however, we do not know how accurate reported kinematic outcomes are because of an inherent source of error known as soft tissue artifact (STA). This error is defined as the movement of skin markers relative to the underlying bone and affects the thigh markers more than the shank. It cannot be filtered out of data after processing because it has a similar frequency content to the movements themselves (~5-10 Hz). Therefore, all reported measures of knee kinematics obtained using optical motion capture include an unknown level of error and can affect clinical and biomechanical interpretations of knee pathologies. This thesis investigated the use of a double calibration technique to improve the accuracy of landmarks tracked in high flexion postures with an anterolaterally located marker cluster in the mid-thigh region. Thirty-three participants performed flatfoot squatting, heels-up squatting, dorsiflexed kneeling and plantarflexed kneeling movements. The position of the functional hip joint center was defined in a standing reference position using a functional calibration trial. This landmark was then simultaneously tracked with the thigh cluster and pelvis cluster during high flexion movements. The landmark tracking with the thigh cluster was referred to as the femoral head center and was obtained in two ways: using either single or double calibration techniques. The landmark tracking with the pelvis, referred to as the hip joint center, was considered the gold standard and thus was the position to which the single and double calibration-based femoral head center positions were compared. Root-mean squared (RMS) error was calculated between the hip joint center and femoral head centers in the global x, y, and z directions. Resultant error (distance between the hip joint center and the femoral head centers) was also determined. The bias and limits of agreement on the resultant error were used to evaluate the accuracy of locating the femoral head center using each calibration technique relative to the location of the hip joint center. Paired t-tests revealed RMS error and resultant error were not significantly lower using the double calibration technique, and the limits of agreement were wider in the double calibration. Data were then separated into percentile groups to evaluate the double calibration technique after controlling for mid-thigh circumference, a subject-specific variable. RMS errors in the global y and z directions and resultant error were significantly greater in the double calibration for the 25th and 50th percentile groups, while there were no significant differences in RMS error nor resultant error between calibration techniques in the 15th and 75th percentile group. Resultant error also increased from 15th to 50th percentile groups and limits of agreement increased with higher percentiles. Due to poor performance of the double calibration despite separating data into mid-thigh percentile groups, the double calibration method was evaluated using a different landmark to predict the position of the femoral head center in the seated calibration posture (as the functional hip joint center trial was only able to be applied to the standing posture). Double calibration using the lateral femoral condyle performed only slightly superior than using the greater trochanter, the landmark initially used in the double calibration method; however single calibration continued to perform best. Therefore, the double calibration was not recommended for use with the studied high flexion movements. The conclusions of this thesis direct future research to evaluate the composition of the thigh (adipose tissue vs muscular tissue) and its relation to STA, and to assess ways to improve palpation of landmarks in high flexion postures (e.g., using fluoroscopy measures). Once accurate palpation is achieved in high flexion, the double calibration technique could be revisited to evaluate its effectiveness to reduce STA.