A reference frame alignment method for the consistent interpretation of kinematic signals

Abstract

Kinematic analysis involves calculating signals from optical/inertial datapoints to represent the relative movement of joint segments. The exact choice of local segment frame orientation and position has been shown to drastically influence the shape and magnitude of the associated kinematic signals, making the consistent interpretation of the underlying motion a challenge [1,2]. Despite attempts to standardise the reporting of these signals [3], a lack of consensus around joint coordinate frame definitions remains. An approach capable of accommodating different analytical methods and reconciling these differences in frame alignment, while ensuring consistent interpretations, is therefore crucial. Given sets of kinematic data, can mathematical optimisation be leveraged to achieve a consistent interpretation of the underlying movement patterns, independent of joint axis definitions? Here, we assess a REference FRame Alignment MEthod (REFRAME) on the in vivo moving-fluoroscopy-based knee kinematics of 10 healthy subjects (5 trials of stair descent each) [4]. Using three methods of defining the flexion/extension axis (cylindrical axis (CA), functional flexion axis (FFA), and transepicondylar axis (TEA)), three different femoral frames were defined for each trial, in addition to a single tibial frame [1]. Rotations of the tibia relative to the femur were calculated, alongside translational positions of the femoral origins in the tibial frame. By implementing REFRAME (as a constrained nonlinear minimisation of ab/adduction and int/external rotation root-mean-square, in addition to all translation variances), local frames were repositioned and reorientated, to derive a set of "REFRAMEd" signals. Fig. 1 - Knee kinematics (rotations[°]: tibia relative to femur; translations[mm]: femur relative to tibia) during a sample stair descent trial, using three different primary axes, before (raw) and after REFRAME. (CA and FFA partially covered by TEA) Download : Download high-res image (294KB)Download : Download full-size image Across all subjects and trials, before REFRAME implementation, the maximum absolute differences between kinematic signals representing the same underlying movement, but derived using different joint axis approaches, reached 1.61° for flexion/extension, 12.00° for ab/adduction, and 12.02° for int/external rotation, in addition to 2.28 mm for mediolateral, 10.60 mm for anteroposterior, and 12.23 mm for proximodistal translations. After REFRAME, maximum differences peaked at 0.78°, 0.08° and 0.08° for flexion/extension, ab/adduction and int/external rotation, respectively; For translations, values peaked at 0.24 mm, 0.10 mm and 0.13 mm in the mediolateral, anteroposterior and proximodistal directions. Moreover, the three signals converged after REFRAME optimisation (Fig1). For each underlying movement pattern, the analysis approach (method of axis definition) affected the characteristics of the kinematic signals. By implementing REFRAME, tibial and femoral frames associated with each signal set were repositioned and reoriented to a common alignment, without requiring knowledge of the original femoral frames' alignment relative to each other. REFRAME thus enables the consistent interpretation of joint kinematics derived using different approaches.