Evaluation of gait function in children with cerebral palsy: New insights into commonly used methods

Abstract

Walking is an essential motor skill for daily living and social participation. The neuromuscular system is vital in achieving support, stability, and progression during walking. Motor disorders, such as cerebral palsy (CP) may cause deviations in gait pattern, asymmetry, and altered muscle activity with excessive muscle co-activation, affecting gait function. In children with CP, impaired gait function may be seen through increased energy cost of walking compared to typically developing children, which in turn may limit activity level and participation. Evaluation of factors related to both typical and pathological gait function is important for clinical decision making. For such, objective and quantitative methods are needed. A 3-dimensional gait analysis (3DGA) describes the gait pattern by providing kinetic, kinematic, and spatiotemporal information. Global measures, such as the gait deviation index (GDI), may be extracted from the 3DGA data, providing an overall score of gait pathology. Surface electromyography (sEMG) provide information about muscle activity and muscle coactivation during walking. However, there are unwanted factors affecting the sEMG signal, and various ways to deal with this, which complicates the interpretation. Moreover, various methods are used to calculate the co-activation index. Energy cost of walking represents the overall gait function and may be presented as total (gross energy cost) or in addition to resting energy expenditure (net energy cost). Gross energy cost is considered more reliable, while net energy cost is reported less affected by between-subject variations in speed and growth-related subject characteristics. However, the effect of the within-subject variation in speed on energy cost is less established. Additionally, to what extent energy cost is affected by a deviating and more asymmetric gait with increased muscle co-activation, is still deficient. Therefore, this thesis aimed to gain further insight into commonly used methods for evaluating gait function in ambulant children with CP. Paper I evaluated how the interpretation of muscle activity and co-activation was affected by the normalisation of the sEMG-amplitudes, and compared two methods of calculating the coactivation index. The findings showed that the overall muscle activity pattern did not change after normalisation, but the between-subject variation was reduced. However, relevant physiological variation may have also been eliminated. The children with CP showed deviations in muscle activity from the typically developing children in different phases of the gait cycle using absolute and normalised amplitudes. Consequently, some of the phases with increased co-activation index deviated between the two methods. But in common they showed that an increase in the muscle co-activation index mostly was attributed reduced muscle activity rather than increased antagonist muscle activity. The two methods of calculating the coactivation index showed similar deviations between the groups, but one method was considered less applicable due to greater between-subject variation and non-normal distribution. Paper II evaluated gross and net energy cost in typically developing children, and the effect of speed and growth-related subject characteristics. The findings showed that gross energy cost was less affected by within-subject variations in speed compared to net energy cost, where an increase in speed showed increased energy cost. Gross energy cost had a strong, negative relation to between-subject variation in speed, age, and body size. Although reduced, these relations were not eliminated for net energy cost, and they followed a concave shape. Paper III evaluated how gross energy cost of walking in children with CP was affected by gait pattern, gait asymmetry, and lower limb muscle co-activation. The findings showed that deviations in gait pattern, reflected through the GDI, in addition to body size, had a strong, positive relation to energy cost. Gait asymmetry and co-activation were not related to energy cost. To summarise, the interpretation of muscle activity and co-activation was affected by normalisation method when evaluating group differences. Thus, using both absolute and normalised amplitudes for a complete interpretation of the sEMG data is recommended. Moreover, when interpreting the co-activation index, the methodological approach and the underlying muscle activity must be considered before drawing conclusions on abnormal coactivation levels and the cause of this. Although normalised to speed and body size, both gross and net energy cost of walking in typically developing children were still affected by those factors. However, gross energy cost may be beneficial when evaluating treatment effect, as improvements in energy cost due to improvements in gait function and consequently speed, may be concealed using net energy cost. Moreover, increasing deviations in gait pattern among children with CP was related to increase in gross energy cost of walking.