Abstract
Preparing children with cerebral palsy prior to gait analysis may be a challenging and time-intensive task, especially when large number of sensors are involved. Collecting minimum number of electromyograms (EMG) and yet providing adequate information for clinical assessment might improve clinical workflow. The main goal of this study was to develop a method to estimate activation patterns of lower limb muscles from EMG measured from a small set of muscles in children with cerebral palsy. We developed and implemented a muscle synergy extrapolation method able to estimate the full set of lower limbs muscle activation patterns from only three experimentally measured EMG. Specifically, we extracted a set of hybrid muscle synergies from muscle activation patterns of children with cerebral palsy and their healthy counterparts. Next, those muscle synergies were used to estimate activation patterns of muscles, which were not initially measured in children with cerebral palsy. Two best combinations with three (medial gastrocnemius, semi membranous, and vastus lateralis) and four (lateral gastrocnemius, semi membranous, sartorius, and vastus medialis) experimental EMG were able to estimate the full set of 10 muscle activation patterns with mean (± standard deviation) variance accounted for of 79.93 (± 9.64)% and 79.15 (± 6.40)%, respectively, using only three muscle synergies. In conclusion, muscle activation patterns of unmeasured muscles in children with cerebral palsy can be estimated from EMG measured from three to four muscles using our muscle synergy extrapolation method. In the future, the proposed muscle synergy-based method could be employed in gait clinics to minimise the required preparation time.
Highlights
Preparing children with cerebral palsy prior to gait analysis may be a challenging and timeintensive task, especially when large number of sensors are involved
In addition to the variance accounted for (VAF), R2, and root mean squared error (RMSE), we introduced another evaluation metric known as Kolmogorov–Smirnov (KS) test[9] to gauge the neural information preserved in the estimated muscles
Mean, and minimum VAF (%) across all children with cerebral palsy (CP) were calculated for each combination of included muscles from which EMGs needs to be recorded (Fig. 2)
Summary
Preparing children with cerebral palsy prior to gait analysis may be a challenging and timeintensive task, especially when large number of sensors are involved. Muscle synergies are extracted from processed electromyograms (EMG), known as muscle activation patterns, using mathematical factorisation m ethods[9] These methods assume recorded muscle activation patterns to be linear combinations of excitation primitives, which represent the temporal profile of muscle synergies[10,11]. The number and location of muscles included in a synergy analysis directly impact the reconstruction e rror[12] It is not known which set of muscles generate muscle synergies that best describe how muscles are activated synchronously during walking in children with CP. The temporal profile of muscle synergies was altered in children with CP after a botulinum toxin type A injection[14] and after selective dorsal rhizotomy[15] The former temporarily inhibits signal from nervous system to block muscle activity[13,14] and the latter is a single event surgical procedure to decrease spasticity through reduction of the excitatory input of dorsal nerve rootlets[16]. studies[14,19,21] suggested that the number of muscle synergies may be difficult to change with an intervention, instead proposing the temporal profile of muscle synergies to be a viable target for treatment of children with CP
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
