Estimation of multivariable dynamic ankle impedance after botulinum toxin injection in children with cerebral palsy

The main object of this study is to quantify the effect of botulinum toxin on the dynamic ankle impedance (ZAnkle) in children with cerebral palsy (CP). The effect was evaluated by comparing the ZAnkle before and one month later the botulinum injection in gastrocnemius muscle. A cohort of six children with CP was involved in the study. All data were gathered by the pediatric version of the Anklebot (PediAnklebot), a backdrivable robot for ankle rehabilitation. The robot applied random torque perturbations to the ankle in both inversion-eversion (IE) and dorsi-plantar flexion (DP) direction. The ankle angular displacements and torques were gathered by encoders and load cells mounted on the PediAnklebot. The ZAnkle was estimated by a Multiple-Input Multiple-Output system identification (MIMO). The ZAnkle was evaluated in three frequency ranges: 0 Hz ≤ f ≤ 2 Hz, 2 Hz ≤ f ≤ 5 Hz, and 5 Hz ≤ f ≤ 8 Hz, addressed as low, mid and high frequency ranges. In each selected range the impedance anisotropy was quantified by the Direction Dependent Map (DDM) and the DDM shape was evaluated via the ZRatio, defined as the ratio between the maximum and the minimum value assumed by ZAnkle. Statistically significant decreases induced by the toxin treatment were found for ZAnkie in IE direction only in high frequency range and in DP direction for the three frequency ranges here examined. Moreover, a significant decrease of ZRatio emerged at mid frequency range. In conclusion, the botulinum toxin injection mainly affected the movement involving gastrocnemius, i.e. the dorsi-plantar flexion, and it did not affect the ankle impedance in inversion-eversion. These findings implied a clear variation of muscle elasticity and can provide an objective and quantitative support in the neurorehabilitation treatments of pediatric population with neurological diseases.