IMU-based assessment of motor control in a population of young subjects with idiopathic scoliosis: preliminary results from the motor-child study

Abstract

Motor development is an adaptive process occurring within a self-organizing system and influenced by several factors. While evidence exists for motor control in severe pathologies (cerebral palsy), minor musculoskeletal conditions, such as scoliosis, have never been explored. Clinical-functional tests used for the analysis of motor control (Movement ABC-2 or TGMD-2) are time-consuming, hindering their applicability to large populations. In recent years wearable Inertial Measurement Units (IMUs) were introduced for low-cost ambulatory quantitative assessment allowing the estimation of temporal gait parameters [1] and of dynamics of motor control through nonlinear metrics [2–4]. The aim of this work was to assess motor control alterations in children affected by idiopathic scoliosis using clinical- functional tests and IMU-based metrics related to motor control. 34 children aged between 6 and 16 years with scoliosis (ATR ≥ 5 ° or Cobb ≥ 10°) were included. They underwent MovementABC-2 for clinical assessment of motor competence. Three tri-axial IMUs (MetamotionR, mbient-Lab, USA) were attached on the lower back and on the shanks [5] of each child. 3D acceleration and angular velocity data were recorded at 200 Hz while the participants walked at self- selected speed back and forth, two times in normal (NW) and once in tandem walking (TW) along a 15 m path. Stride (StrideT, seconds), stance (StanceT, % of StrideT), and double support time (DS, % of StrideT) were calculated from foot contact events. Temporal parameter variability was calculated as standard deviation. MSE (τ from 1 to 6), related to complexity and automaticity, and RQA, related to pattern regularity, were calculated on trunk acceleration data along the 3 directions. Data were compared to those of matched healthy peers. Movement ABC-2 test: 12 patients received scores above the 50th percentile (35.3%) and 22 below (64.7%). No significant correlation was observed with the degree of scoliosis (Cobb angle) or the degrees of torsion (ATR). Temporal parameters: subjects resulted slower during NW (longer StrideT), with larger dispersion of StanceT and DS than healthy peers. During TW, subjects above 12 years of age resulted slower, with longer StanceT and longer DS for increasing age as compared to their healthy peers. Non-linear metrics: subjects showed larger MSE during NW for lower values of τ (1 to 3) in the ML and V direction and reduced RQA in AP direction for increasing age as compared to healthy peers; during TW, reduced MSE for lower values of τ (1 to 3) in the V and AP direction and increased RQA in AP direction. Children affected by idiopathic scoliosis showed alterations of motor control. In particular, temporal parameters highlighted a slower and less stable gait, more evident for increasing age as related to TW. Lower stability is also related to a less mature motor control with a delay in the development of the automaticity characterizing healthy peers, and a reduced development of the motor complexity required for more demanding locomotor tasks. These preliminary data suggest that idiopathic scoliosis affects motor control in the developing population. IMU-derived metrics demonstrated their effectiveness in complementing clinical assessment.